The Structural Biodegradation of Graphene Oxide in Vivo

The Structural Biodegradation of Graphene Oxide In Vivo

A thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Biology, Medicine & Health

Leon Newman

School of Health Sciences, Division of Pharmacy and Optometry

2017

Table of Contents List of Figures ...... 5

List of Tables ...... 10

Abstract ...... 12

Declaration ...... 13

Acknowledgements ...... 14

Rationale for the alternative thesis format ...... 15

List of abbreviations ...... 16

Chapter 1 ...... 18

1. Introduction...... 18

1.1. Introduction and overview of thesis contents...... 18

1.2. References ...... 23

Chapter 2 ...... 24

2. Understanding the degradability of sp2 hybridised carbon nanomaterials for their applications in biomedicine ...... 24

2.1. Statement...... 24

2.2. Abstract...... 25

2.3. Introduction...... 26

2.3.1. sp2 hybridised CNMs...... 27

2.3.2. Fullerenes...... 27

2.3.3. Carbon nanotubes...... 28

2.3.4. Carbon nanohorns...... 30

2.3.5. Graphene...... 32

2.3.6. Functionalised derivatives...... 33

2.4. Degradation and the use of CNMs in biomedicine...... 35

2.5. Chemically mediated degradation...... 36

2.5.1. Overview of chemical mediated CNM degradation and the lessons learnt...... 42

2.6. Enzyme catalysed biodegradation...... 46

2.6.1. The peroxidase enzyme superfamily...... 46

2.6.2. Horse radish peroxidase...... 46

2.6.3. Myeloperoxidase...... 51

2.6.4. Other enzymes...... 54

2.6.5. Overview of enzyme mediated CNM biodegradation and the lessons learnt...... 57

2.7. Biodegradation mediated by cells...... 61

2.7.1. Neutrophils...... 61

2.7.2. Macrophages...... 65

2.7.3. Microglia...... 67

2.7.4. Overview of cell mediated CNM biodegradation and the lessons learnt. 71

2.8. Biodegradation of CNMs in vivo...... 76

2.8.1. Biodegradation of CNMs in the lungs...... 76

2.8.2. Biodegradation of CNMs in other tissues...... 79

2.8.3. Overview of in vivo mediated CNM biodegradation and the lessons learnt...... 80

2.8.4. Summary of the factors that can be used to modulate CNMs biodegradation relevant to biomedicine...... 84

2.9. Products of degradation and their health effects...... 84

2.10. Biodegradation requirements for CNM based biomedical applications and beyond...... 87

2.10.1. Drug delivery...... 87

2.10.2. Tissue engineering...... 91

2.10.3. Diagnostics and monitoring...... 94

2.10.4. Summary of the need to consider CNM biodegradation in relation to their biomedical applications...... 96

2.11. Conclusion...... 97

2.12. References...... 99

Chapter 3 ...... 115

3. Hypothesis, Aims and Objectives ...... 115

Chapter 4 ...... 117

4. Hypochlorite Degrades 2D Graphene Oxide Sheets Faster than 1D Oxidised Carbon Nanotubes and Nanohorns ...... 117

4.1. Statement...... 117

4.2. Abstract...... 118

4.3. Introduction...... 119

4.4. Results...... 121

4.5. Discussion...... 127

4.6. Experimental...... 135

4.7. Acknowledgements...... 138

4.8. Supplementary information ...... 139

4.9. References ...... 148

Chapter 5 ...... 152

5. Splenic capture and in vivo intracellular biodegradation of thin, biological-grade graphene oxide sheets ...... 152

5.1. Statement...... 152

5.2. Abstract...... 153

5.3. Introduction...... 154

5.4. Results...... 156

5.5. Discussion...... 169

5.6. Experimental...... 178

5.7. Acknowledgements...... 189

5.8. Supplementary information...... 190

5.9. References...... 207

Chapter 6 ...... 213

6. Nose to Brain Translocation and Biodegradation of Graphene Oxide Nanosheets ...... 213

6.1. Statement...... 213

6.2. Abstract...... 214

6.3. Introduction...... 214

6.4. Results...... 218

6.5. Discussion...... 235

6.6. Experimental...... 245

6.7. Acknowledgments...... 252

6.8. Supplementary Information...... 253

6.9. References ...... 264

Chapter 7 ...... 273

7. Summary and Final Remarks ...... 273

7.1. Summary...... 273

7.2. Conclusions and further work...... 277

7.3. References...... 278

List of Figures

Figure 2.1 An illustration of the typical structure of a truncated icosahedron C60 fullerene. .. 28 Figure 2.2 An illustration of the structure of MWNTs...... 29 Figure 2.3 An illustration of the different forms of SWNT that can be derived from a 2D sheet of graphene (zig zag, arm chair and chiral)...... 30 Figure 2.4 An illustration of the structures of CNHs and dahlia-like superstructures...... 32 Figure 2.5 An illustration of graphene as the archetypal member of the family of CNMs. .... 33 Figure 2.6 The structure of GO...... 34 Figure 2.7 A schematic diagram demonstrating the facile degradability of GO via the photo- Fenton reaction...... 39 Figure 2.8 Binding of human serum albumin (HSA) to SWNTs reduces the kinetics of NaClO-mediated degradation in vitro...... 41

Figure 2.9 A summary of the factors that can influence the kinetics of chemically mediated CNM degradation...... 43 Figure 2.10 TEM images demonstrating how the enzymatic biodegradation of MWNTs can be altered via the conjugation of specific substrates...... 49 Figure 2.11 The condition specific biodegradation of disulphide-linked GO-SS-PEG...... 51 Figure 2.12 TEM images of hMPO-mediated degradation which evidences that MPO mediated degradation of GO sheets can be regulated by the dispersibility and aggregation state of the material...... 54 Figure 2.13 A summary of the factors that can influence the kinetics of enzymatically mediated CNM biodegradation...... 58 Figure 2.14 Proposed mechanisms for IgG coating-dependent degradation of SWNTs in neutrophils...... 63 Figure 2.15 NET MPO mediated biodegradation of SWNTs...... 64 Figure 2.16 Raman spectroscopy of microglia cell culture monolayers after exposure to different functionalised MWNTs...... 70 Figure 2.17. Summary of the factors that can influence the kinetics of cell mediated CNM biodegradation. 72 Figure 2.18. Raman spectroscopic evaluation of ‘‘oxidative’’ defects in SWNT present in the lungs of wild-type (w/t) and MPO knocked-out (k/o) mice at days 1 and 28 post exposure using single point Raman spectroscopy or Raman mapping of different areas within the tissue samples...... 78 Figure 2.19 A summary of the factors that can influence the kinetics of CNM biodegradation in vivo...... 81 Figure 2.20 An illustration of the various parameters that can affect the biodegradation kinetics of different CNMs in view of biomedical applicability...... 84 Figure 2.21 An illustration of the different biodegradation kinetic requirements of CNMs when incorporated into various biomedical strategies...... 97

Figure 4.1 Characterisation of the starting oxidised carbon nanomaterials...... 122 Figure 4.2 Optical changes of oxidised carbon nanomaterials over time in 1% NaClO...... 123 Figure 4.3 Evolution of Raman spectra of oxidised carbon nanomaterials...... 125 Figure 4.4 Representative observations of ultra-structural changes in oxidised carbon nanomaterials exposed to 1 % NaClO as detected by TEM (left) and AFM (right) over time...... 126 Figure 4.5 Schematic representation of the degradation processes...... 134

Figure 4.S1 Analysis of functionalities present on the surface of starting oxidised carbon nanomaterials using XPS. 139 Figure 4.S2 Chemical composition of the starting oxidised carbon nanomaterials as analysed by TGA. 140 Figure 4.S3 Evolution of the UV-vis absorption spectra of oxidised carbon nanomaterials in 1

% NaClO and H2O over time. 141 Figure 4.S4 Representative observations of ultra-structural changes of oxidised carbon

nanomaterials incubated in H2O as detected by TEM over time. 142 Figure 4.S5 TEM images of the wall exfoliation in OxMWNT when incubated in 1 % NaClO. 143 Figure 4.S6 Evolution of Raman spectra of GO over time. 144 Figure 4.S7 Evolution of Raman spectra of OxMWNT over time. 145 Figure 4.S8 Evolution of Raman spectra of OxNH over time. 146 Figure 4.S9 Evolution of ATR FTIR spectra of oxidised carbon nanomaterials over time. 147

Figure 5.1 Physicochemical characterisation of the starting GO material...... 157 Figure 5.2 Effect of GO on spleen structure...... 159 Figure 5.3 Effect of GO on spleen haematological function, after 24 h and 1 month of injection of GO at different concentrations (2.5, 5, 10 mg/ml) compared to control mice, Dex 5% (negative control) and LPS 5 mg/kg injected (positive control)...... 161 Figure 5.5 Effect of GO on spleen immunological function, after 24 h and 1 month of injection of GO at different concentrations (2.5, 5, 10 mg/ml) compared to control mice, Dex 5% (negative control) and LPS 5 mg/kg injected (positive control)...... 163 Figure 5.6 Splenic localisation and abundance of GO at different time points following the administration of 7.5 mg/kg GO...... 165 Figure 5.7 Overview of the splenic degradation of GO over 9 months following its intravenous administration (7.5 mg/kg)...... 168

Figure 5.S1 Physicochemical characterisation of starting GO material...... 192 Figure 5.S2 SAED aperture micrographs of the region of interest corresponding to the electron diffractogram in Figure 5.1. The scale bar represents 200 nm...... 193 Figure 5.S3 The gross appearance of the spleens of C57BL/6 mice 24 h and 1 month after injection with GO at different concentrations (2.5, 5 and 10 mg/mL), compared to those of control mice (mice injected with Dex 5%)...... 193 Figure 5.S4 Murine spleen sections that were stained with the TUNEL stain at 24 and 1 month post GO administration with at different concentrations (2.5, 5, 10 mg/kg) or Dex 5%...... 194 Figure 5.S5 Apoptotic cells density as per TUNEL assay at 24 and 1 month post-injection with GO at different concentrations (2.5, 5, 10 mg/kg) or Dex 5%...... 195 Figure 5.S6 Representative flow cytometry plots outlining the CD4 and CD8 cell counts of mice 24 h and 1 month after administration of GO at different concentrations (2.5, 5 and 10 mg/mL), compared to those obtained for control mice, Dex 5% ( negative control and LPS-injected mice at 24 h (positive control)...... 196 Figure 5.S7 Growth curve of mice injected with GO (7.5 mg/kg) or Dex 5% (control) over 9 months...... 197 Figure 5.S8 Murine spleen sections stained with TUNEL stain at different time points (Day 1, 30 and 270) after GO intravenous administration (7.5 mg/kg)...... 198 Figure 5.S9 Cellular distribution of splenic GO at different time points following GO administration (7.5 mg/kg)...... 199 Figure 5.S10. The evolution of the average Raman spectra of splenic captured GO over 9 months...... 200 Figure 5.S11 Representative H & E stained splenic sections of mice at different time points following GO administration (7.5 mg/kg)...... 201 Figure 5.S12 TEM micrographs of GO sequestered within the vesicular compartments of murine marginal zone splenocytes at different time points following administration of 7.5 mg/kg GO...... 202 Figure 5.S13 Representative electron diffractograms of intracellularly entrapped GO material within the spleens of mice at different time points following administration of GO (7.5 mg/kg), with the corresponding intensity line profiles...... 203 Figure 5.S14 Holes seen within intracellularly entrapped GO sheets, which are indicative of an early transitional state of GO as its biodegradation proceeds...... 204 Figure 5.S15 Physicochemical characterisation of GO material after 9 months incubation in water at 37 ̊C in darkness...... 205

Figure 5.S16 SAED aperture micrographs of the region of interest corresponding to the electron diffractogram in Figure 5.1...... 206

Figure 6.1 Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA functionalisation...... 220 Figure 6.2 Indium chelation efficiency and stability of GO-DOTA[In] conjugates...... 222 Figure 6.3 Brain-specific biodistribution of GO sheets of different lateral dimensions 24 hours following a single intranasal administration...... 225 Figure 6.4. Raman-based brain biodistribution of GO sheets of different lateral dimensions 24 hours following a single intranasal administration...... 227 Figure 6.5 Qualitative and semi-quantitative location of GO 24 h following a single intranasal administration...... 229 Figure 6.6 Evolution of the biodistribution profile of us-GO at days 1 and 7 following a single intranasal administration, assessed by ICP-MS...... 231 Figure 6.7 Evolution of the biodistribution profile of us-GO at day 1 and 7 following a single intranasal administration, assessed by Raman spectroscopy...... 232 Figure 6.8 Biodegradation of us-GO over 28 days following a single intranasal administration...... 234

Figure 6.S1 Schematic illustrating the epoxide ring-opening reaction mechanism which we used to functionalise GO sheets (l-GO, s-GO or us-GO) mechanism with DOTA. 254 Figure 6.S2 Dimensional analysis of starting and DOTA functionalised l-GO, s-GO and us- GO. 255 Figure 6.S3 Spectroscopic features of l-GO, s-GO and us-GO materials and their respective DOTA functionalised derivatives as analysed using Raman spectroscopy and ATR FT- IR. 256 Figure 6.S4 SPECT/CT-based qualitative biodistribution at early time points (30 mins and 3 hours post administration). 257 Figure 6.S5 The presence of GO in the extracted nasal cavities of mice after 24 hours following treatment with unlabelled l-GO, s-GO and us-GO. 257 Figure 6.S6 Raman spectra used as reference for Raman-based correlative mapping of brain sections as shown in Figure 6.4. 258 Figure 6.S7 Bright field images showing regions of interest where confocal Raman maps were acquired. 259

Figure 6.S8 Changes in the brain biodistribution profile of us-GO at day 1 and 7 following a single intranasal administration. 260 Figure 6.S9 Presence of GO in the brain decreases over 28 days following intranasal instillation. 261 Figure 6.S10 Degradation of us-GO 28 days following multiple intranasal administrations.262 Figure 6.S11 Growth curve of C57BL/6 mice administered instranasally with us-GO on a single occasion or on multiple occasions. 263

List of Tables Table 2. 1 A list of the literature studies that have specifically focused on the chemically mediated degradation of CNMs...... 44 Table 2.2 A list of the literature studies that have specifically focused on the enzyme mediated degradation of CNMs...... 59 Table 2.3 A list of literature studies that have specifically focused on the cell mediated degradation of CNMs...... 73 Table 2.4 A list of literature studies that have specifically focused on the in vivo degradability of CNMs...... 82

Table 5.S1 Statistics table, showing p values, describing the differences in the percentage of marginal zone macrophages in which a Raman signature characteristic of GO (corresponding to Figure 5.4C) could be acquired between the respective time points following administration with GO (7.5 mg/kg)...... 190 Table 5.S2 Statistics table, showing p values, describing the differences between the apoptotic cell density (apoptotic cells / mm2) detected in spleen sections of mice treated with GO (7.5 mg /kg) treated or Dex 5% at different time points...... 190 Table 5.S3 Statistics table (corresponding to Figure 5.5B), showing p values, illustrating the statistical differences in the number of dark areas per marginal zone between different time points following the administration of 7.5 mg/kg GO...... 191 Table 5.S4 Statistics table, showing p values, describing the differences in the I(D)/ I(G) ratios between different time points following the intravenous administration of 7.5 mg/kg of GO...... 191

Table 6.S1 Physicochemical analysis of non-functionalised l-GO, s-GO and us-GO materials used for unlabelled Raman-based biodistribution and degradation studies. The lateral dimension and thickness values represent ranges i.e. the largest and smallest values detected. All percentages derived from XPS measurements represent atomic percentages...... 253

Total Word Count: 78,637

Abstract University of Manchester

Name: Leon Newman

Degree Title: Doctor of Philosophy

Thesis title: The Structural Degradation of Graphene Oxide In Vivo

Date: 2018

In the field of biomedicine, the oxidised derivative of graphene, graphene oxide (GO), has been intensively researched for applications such as drug delivery, tissue engineering as well as medical imaging. GO has also shown applicability to areas of material science including in the design of anticorrosion coatings and desalinisation membranes. Consequently, there is a risk of the intended or unintended exposures of humans. It is crucial that the biological interactions and fate of GO are revealed. The aim of this thesis is to help in the revealing of this information. Herein, the in vivo biodegradability or susceptibility of GO sheets to be structurally destroyed is explored. Currently, research examining the biodegradability of GO sheets is limited to a few studies.

This thesis starts with an in vitro study to characterise the degradative changes that occur in GO sheets under a defined oxidative influence. The author then describes the interrogation of the biodegradability of GO sheets in the spleen and the brain following intravenous and intranasal exposures respectively, in C57BL/6 mice. In order to reach conclusions, various analytical techniques were used, including microscopic (transmission electron microscopy, atomic force microscopy, light microscopy), spectroscopic (Raman spectroscopy), spectrometric (inductively coupled plasma mass spectrometry) and histological techniques (immunostaining and standard hematoxylin and eosin staining) as well as correlative techniques (immunostaining followed by Raman spectroscopy on the stained cells or tissue sections). The data indicate that under in vivo conditions in the C57BL/6 mice strain, GO nanosheets undergo biodegradative changes mediated by macrophage lineage cells (marginal zone macrophages in the spleen or microglia in the brain). The toxicological effects of GO on the splenic structure and function were also explored. The results indicate that the intravenously administered thoroughly characterised GO sheets did not jeopardise the splenic structure or function. Regarding the nose-to-brain focused study, the effect of GO sheet lateral dimension on the extent of translocation was investigated. GO sheet lateral dimension was inversely related to extent to which GO sheets could translocate from the nose to the brain following intranasal administration, with the GO sheets of lowest lateral dimensions translocating to the greatest extent. A fraction of the translocated GO sheets was seen to diffuse to distal regions of the brain past the initially encountered olfactory bulb.

The findings of this work have implications that are relevant to potential clinical applications, where biodegradability is a critical parameter. Moreover, it is crucial for helping to complete the toxicological profile of GO. Nonetheless, further studies are necessary to understand the molecular mechanisms by which the observed in vivo biodegradation occurs and the identities of the by-products that are generated as a result of this process.

Declaration

Declaration No portion of the work referred to in the thesis has been submitted in support of an application for another degree or qualification of this or any other university or other institute of learning with the exception of Chapter 6, which was in part presented as part of my colleague and co-author’s PhD thesis (Dhifaf Jasim, University of Manchester, 2016).

Copyright statement i. The author of this thesis (including any appendices and/or schedules to this thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he has given The University of Manchester certain rights to use such Copyright, including for administrative purposes. ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic copy, may be made only in accordance with the Copyright, Designs and Patents Act 1988 (as amended) and regulations issued under it or, where appropriate, in accordance with licensing agreements which the University has from time to time. This page must form part of any such copies made. iii. The ownership of certain Copyright, patents, designs, trademarks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright works in the thesis, for example graphs and tables (“Reproductions”), which may be described in this thesis, may not be owned by the author and may be owned by third parties. Such Intellectual Property and Reproductions cannot and must not be made available for use without the prior written permission of the owner(s) of the relevant Intellectual Property and/or Reproductions.

iv. Further information on the conditions under which disclosure, publication and commercialisation of this thesis, the Copyright and any Intellectual Property and/or Reproductions described in it may take place is available in the University IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=2442 0), in any relevant Thesis restriction declarations deposited in the University Library, The University Library’s regulations (see http://www.library.manchester.ac.uk/about/regulations/) and in The University’s policy on Presentation of Theses

Acknowledgements

I would like to give a huge thanks to Prof Kostas Kostarelos and Dr Cyrill Bussy for giving me the opportunity to pursue a PhD in the Nanomedicine Lab. I would not have been able to complete this PhD if it was not for their advice, support, patience, and their inspirational knowledge and expertise. Both of my supervisors have been and are great sources of knowledge and wisdom for me with regards to science, communication, and career advice and an array of other subjects. I thank you both for all your time during our meetings where we discussed my project, as well as some of the other random and funny topics that our conversations occasionally used to drift off to. Both of you have made my PhD a very fond and excellent time in my life that I am sure I will regularly look back to in the future.

I would also like to express my gratitude to all those who have worked closely with me during these last few years. I would like specifically mention Dhifaf Jasim, Filipe Rodrigues, Eric Prestat and Neus Lozano for all the invaluable and important contributions they have made to my work. I thank all other members of the Nanomedicine Lab, past and present, who have also shared with me my experience. I also thank all the members of the Laboratory of Polymers and Biomaterials especially Prof Nicola Tirelli, Christopher Yusef Leon Valdivieso, Richard D’arcy and Alfredo Gravagnuolo for allowing access and for many helpful discussions regarding FT-IR. You all have made my time in lab a very memorable experience.

I thank Prof Alberto Bianco and Isabella Vacchi from the CNRS for their fruitful discussions and for providing me the functionalised GO materials that were used in my experiments. Also I would like to thank Dr Aleksandr Mironov and Ms. Samantha Forbes for training and supporting me on the TEM. I am also very grateful to Dr Nigel Hodson for training me in AFM. Also I thank Mr Peter Walker who trained me and discussed with me about all aspects of histology. Without all of your expertise this thesis would not be complete.

Finally, I would like to thank my family especially my Mum and Dad and my brothers, Calvin and Jason, for all their support during my PhD. In particular I thank my best friend, fiancé and life partner Jashoda Patel who offered endless support right from the start on this long journey.

Rationale for the alternative thesis format

There are several reasons for which this thesis is presented in the journal format. Firstly, it has provided me with good practice and valuable experience in the presentation of my research in a publishable format, which will be very useful for my future career. Secondly, during my doctoral studies I have worked on various projects that could contribute the literature. By presenting my research as papers, if possible, my work can be published shortly after completion of the PhD.

List of abbreviations

Abbreviation Expanded text 1D One-dimensional 2D Two-dimensional 3D Three-dimensional Å Angstrom AFM Atomic force microscopy BBB Blood brain barrier BF-TEM Bright field – Transmission electron microscopy BSA Bovine serum albumin ClO- Hypochlorite CNH Carbon nanohorn CNM sp2 hybridised carbon nanomaterial CNS Central nervous system CNT Carbon nanotube Dex Dextrose DOTA Tetraazacyclododecane-1,4,7,10-tetraacetic acid EPO Eosinophil Peroxidase fGO Functionalised graphene oxide FT-IR Fourier transform infrared spectroscopy GBM Graphene-based materials GO Graphene oxide h Hour HClO Hypochlorous acid HRP Horse radish peroxidase HSA Human serum albumin i.v. Intravenous ICP-MS Inductively coupled plasma mass spectrometry In-111 Indium - 111 (radioactive) In-115 Indium – 115 (non - radioactive or cold) iNOS Inducible nitrogen oxide synthase l-GO Large graphene oxide sheets LiP Lignin peroxidase LPO Lactoperoxidase

LPS Lipopolysaccharide min Minute mL Millilitre MPO Myeloperoxidase MWNT Multiwalled carbon nanotube MZM Marginal zone macrophage NaClO Sodium hypochlorite NADPH Oxidase Nicotinamide adenine dinucleotide phosphate-oxidase nm Nano-meter Ox Oxidised / Carboxylated OxNH Oxidised carbon nanohorns PEG Poly ethylene glycol PSF Phagolysosomal simulant fluid RBC Red blood cells SAED Single area electron diffraction SEM Scanning electron microscopy s-GO Small graphene oxide sheets SOD Superoxide dismutase SPECT / CT Single-photon emission computed tomography SWNT Single walled Carbon nanotube TEM Transmission electron microscopy TLC Thin layer chromatography TUNEL Terminal deoxynucleotidyl transferase (TdT) dUTP Nick-End Labeling us-GO Ultra-small graphene oxide sheets XO Xanthine peroxidase XPS X-Ray Photoelectron Spectroscopy X-Ray CT X-Ray CT μg Microgram μm Micrometre

Chapter 1

1. Introduction

1.1. Introduction and overview of thesis contents.

This thesis has been designed to focus on the elucidation of the in vivo biodegradability of graphene oxide nanosheets (GO). This material is the oxidised derivative of graphene and has been attributed with various outstanding properties including high surface area, ease of chemical functionalisation, as well as unique optical properties just to name a few1. The investigations contained within this thesis have been conducted in order to help better understand the fate of this unique nanomaterial following its intentional or unintentional exposure to mammalian organisms. Such research is highly relevant in view of the progressive development of GO towards various biomedical and non-biomedical technologies alike.

To begin, Chapter 2 presents a comprehensive review of the published sp2 hybridised carbon nanomaterial (CNM) biodegradation literature. To garner a full understanding of the biodegradability, several CNMs were incorporated into the discussion besides graphene, including fullerenes, carbon nanotubes, carbon nanohorns as well as various functional derivatives of these carbon nanomaterials. As part of the introduction, the unique structural features of CNMs are discussed. This topic runs through all the subsequent chapters and is necessary to fully understand CNMs biodegradation. The essay then elaborates on the different systems that have been shown to effect the degradation of different CNMs. A particular emphasis is put on the factors that can influence the kinetics of CNM degradation mediated by systems of increasing complexity (isolated chemicals all the way to living organisms). Fittingly, the degradation of CNMs mediated by non-living isolated chemicals and enzymes is first discussed. The focus is then redirected to living systems including cells and mammalian organisms, which are the main biodegradation mediating systems investigated in the majority of the experimental work described herein.

Based on the literature, it is found that factors such as the number of layers (or walls), the oxidation degree, the presence of defects, the dispersibility state as well as molecules that coat the CNM surfaces can influence CNMs biodegradation kinetics. These parameters should therefore be considered crucial when optimising CNMs for different biomedical

18 applications, where the control of biodegradation is critical. The literature indicates that as degradative mediating systems become more complex, i.e. from non-living isolated chemicals and enzymes to living cells and multicellular mammalian organisms, the responses of the system to the materials properties become more significant. Such responses can for example accelerate or retard the CNM biodegradation kinetics. This means that studies that focus purely on enzymatic or chemical mediated degradation to predict in vivo mechanisms can miss details that could jeopardise their predictive value. In order to understand biodegradation in the context of possible biomedical applications or in the event of unintended exposures, in vivo studies are required. The aims of Chapter 5 and 6 is to address the in vivo biodegradation of GO. Nevertheless, the value of isolated chemical and enzymatic degradation is not forgotten. These studies can offer key insight into the materials behaviour and reactivity under different defined conditions. Though they are not all-encompassing models, they can provide a starting point for understanding degradability under more undefined conditions. For this reason, the experimental work begins with a series of experiments which seeks to interrogate the degradative changes in GO compared to other related oxidised CNMs of different geometries, which is currently lacking in the literature (Chapter 5).

The review ends with an illustration of the necessity of a firm understanding of CNM degradation and its associated kinetics for successful development of GO in biomedicine. We use case studies to illustrate the way in which CNMs degradation must be carefully tuned in differential manners depending on the target biomedical application of these fascinating materials.

In Chapter 3, presents the overarching hypothesis and aims that runs through the work presented herein. The supposition in general, is that under in vivo conditions, GO will be subject to influences that will ultimately result in its progressive structural biodegradation.

Chapter 4 marks the commencement of our experimental work. Here, a simple strategy is employed to degrade GO, using sodium hypochlorite. The degradative changes that occur in GO in vitro when exposed to hypochlorite are characterised and examined. The highly oxidative hypochlorite was chosen due to its relevance to cellular processes, being the product of some mammalian enzyme-catalysed reactions2,3. Moreover, NaClO is also currently used in domestic and industrial processes in sanitation, water purification as well as industrial bleaching processes4. We monitored the oxidative degradation of GO with various spectroscopic and microscopic characterisation techniques. Particularly, it was important to highlight and understand the specific evolution of the material over time as

19 demonstrated by Raman spectroscopy as well as transmission electron microscopy (TEM) based observations. These are two of the key techniques described in the literature to monitor the in vitro degradation of nanomaterials. In addition to this, we supplement our observations with atomic force microscopy in a correlative fashion with TEM. ATR FT-IR is used to confirm that the changes observed were associated with increased oxidation of the materials over time. In order to fill a knowledge gap in the literature, the degradation of GO to that of two other CNMs, namely 1D oxidised carbon nanotubes and 1D oxidised carbon nanohorns of different geometries, under the same conditions are compared and contrasted. This allowed the gleaning of a better perspective of the morphological changes that were occurring. With this increased understanding, the main aims of the thesis namely, to characterise the degradation of GO sheets under in vivo conditions, could be approached in a more informed and directed manner. .

The Nanomedicine Lab has previously published evidence to suggest that after the intravenous administration of labelled GO sheets, a large fraction of the material undergoes urinary excretion5,6. A small but significant fraction of labelled GO however remains within the body and accumulates within the spleen over 24 hours5. The purpose of Chapter 5 therefore was to confirm the splenic bioaccumulation of unlabelled intravenously administered thoroughly characterised GO sheets, its toxicological significance and the long term fate of the material over 9 months. it was of particular interest to understand the spleen’s capacity to biodegrade the GO over time. Using an array of techniques that combined the fields of the biomedical and material sciences, these queries were addressed. These studies confirmed the splenic accumulation of GO. Moreover, the related question: ‘does the spleen - accumulated GO materials induce detectable splenic pathological consequences?’ was addressed. In order to answer this question, an incremental dose escalation and interrogation of the materials influence on the spleen at a structural and functional level for up to 1 month following the GO administration was conducted by Dr Dhifaf Jasim.

The study was then extended using confocal Raman spectroscopy based mapping, and immuno-fluorescence and Raman combinations to determine the sub-splenic distribution of the material at different time points following administration. It was identified that the material initially accumulates within a sub population of splenocytes. Overtime the material becomes increasingly difficult to detect and where it is detectable, the spectral integrity was increasingly attenuated. To explain this, Raman spectroscopy was employed in a more refined manner to detect any characteristic spectral changes alike that described in Chapter 4. This was to identify if the decreasing presence of spleen accumulated GO could be

20 ascribed at least in part to the materials in vivo biodegradation. The higher resolution and direct imaging of the material within the tissue over time was conducted with TEM; this experimental approach taken to yield more information and to confirm the Raman spectroscopic findings. Furthermore, a protocol was optimised that allowed the acquisition of the electron diffraction pattern of GO in tissues in parallel over time. This was in order to help identify the observed materials as GO, and conclude with higher resolution whether there was a gradual loss of the materials crystallinity over time and hence whether the splenic biodegradation of GO was occurring.

The spleen is one of an entire plethora of intricate in vivo physiological organs, and the study of GO biodegradability in the context of this organ is highly relevant to biomedically related exposures. In Chapter 6 therefore, we chose to broaden our work to encompass the exploration of the biodegradability of GO sheets in a completely different physiological in vivo context. We explored the fate of the material within the brain, which is part of the central nervous system (CNS) unlike the spleen which is an integral part of the lymphoid system. In addition to being a direct route to the lungs, the nasal cavities present unique nose-to-brain translocation pathways. These pathways have been shown to be highly relevant for brain exposure following the intranasal administration of various nanoparticles7-9. However despite this, its relevance to CNM exposure is currently undetermined. To add more value and relevance to the study, we chose to administer the materials via the intra-nasal route, which is relevant to unintended occupational and private exposures. It should be noted that the exposure limits are unknown. Any derived results are merely proof of concept and not representative of real world exposures.

At this point of time in the Nanomedicine Lab, the GO synthetic procedures had been improved to enable the production of batches of GO that were composed of sheets of controlled large (l), small (s) or ultra-small (us) lateral dimensions. The author took advantage of this and administered labelled and unlabelled GO sheets that differed in terms of their lateral dimensions to individual groups of experimental mice. The biodistribution of the materials in the brain using spectroscopic (Raman spectroscopy) and spectrometric (ICP- MS) techniques were probed. The results indicate that GO sheets undergo nose-to- brain translocation in a size-dependent manner. The smallest size category of GO sheets (us-GO, 30 – 500 nm) was shown to gain the greatest access to the brain in terms of quantity as well as brain coverage. The absolute quantity of us-GO that underwent nose-to- brain translocation however was approximately 0.01 % of the administered dose. Most of the material that did translocate however accumulated within the olfactory bulb, which lies just above the nasal cavity. Strong evidence was acquired that suggested the further

21 translocation of usGO sheets to more distal brain regions including the cerebellum. The author of this thesis was interested in several aspects of this translocation before the materials biodegradability was studied. Firstly, once in the brain, did this diminutive quantity of us-GO sheets reside with close association to any specific cells? Secondly, did the amount of material persist within the brain over time or was there evidence to suggest a removal mechanism? In order to answer these questions, we optimised Raman based and mass spectrometry based protocols to analyse the presence of labelled and unlabelled materials in brain tissue. It was found that the quantity of us-GO was maintained in the brain over a week. To identify if this retained material was biopersistent or could undergo a biodegradative process, Raman spectroscopy was to probe the materials crystallinity up to a month after single as well as and multiple intranasal administrations. This was conducted in a similar manner to that described in Chapter 4 and 5.

By combining the findings of Chapter 5 and 6, and armed with the information and understanding gained in Chapter 4, the in vivo biodegradation of GO in a lymphoid organ as well as in the central nervous system was addressed. The way in which the studies contained within this thesis were conducted allowed the addressing of other associated and relevant questions including the materials cellular biodistribution following administration. Moreover studies were conducted that addressed the biodegradability of GO after the administration of the material via routes that simulated both intended and unintended exposures. In this way the work contained herein becomes more relevant, considering the scientifically and technologically fruitful future predicted for GO materials and their consequent increasing probability for human exposure. All though as stressed before the exposure levels, in different environments and as graphene technology advances, should be determined.

This last investigation concludes the experimental work described in this thesis. Fittingly in Chapter 7, a reflection, a summary and the provision of ideas for future work is described that could be conducted to further the science portrayed herein.

The referencing style for the Chapter 1, 4, 5, and 7 correspond to Nature guidelines, while for Chapters 2 and 6 the referencing style matches the Royal Society of Chemistry journal guidelines, with the name of the article appearing at the end of the reference for the solely for the convenience of the reader. The referencing style has been completed in this manner to facilitate the submission of the relevant chapters to the respective journals following the completion of the PhD.

1.2. References

1 Dreyer, D. R., Park, S., Bielawski, C. W. & Ruoff, R. S. The chemistry of graphene oxide. Chem. Soc. Rev. 39, 228-240 (2010). 2 Vlasova, II, Sokolov, A. V., Chekanov, A. V., Kostevich, V. A. & Vasil'ev, V. B. Myeloperoxidase-induced biodegradation of single-walled carbon nanotubes is mediated by hypochlorite. Bioorg Khim 37, 510-521 (2011). 3 Kotchey, G. P. et al. A Natural Vanishing Act: The enzyme-catalyzed degradation of carbon nanomaterials. Acc. Chem. Res. 45, 1770-1781 (2012). 4 Fukuzaki, S. Mechanisms of actions of sodium hypochlorite in cleaning and disinfection processes. Biocontrol Sci. Technol. 11, 147-157 (2006). 5 Jasim, D. A., Menard-Moyon, C., Begin, D., Bianco, A. & Kostarelos, K. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. Chem. Sci.6, 3952-3964 (2015). 6 Jasim, D. A. et al. The effects of extensive glomerular filtration of thin graphene oxide sheets on kidney physiology. ACS Nano 10, 10753-10767 (2016). 7 Illum, L. Is nose-to-brain transport of drugs in man a reality? J. Pharm. Pharmacol. 56, 3-17 (2004). 8 Illum, L. Transport of drugs from the nasal cavity to the central nervous system. Eur. J. Pharm. Sci. 11, 1-18 (2000). 9 Oberdörster, G., Elder, A. & Rinderknecht, A. Nanoparticles and the Brain: Cause for Concern? J. Nanosci. Nanotechnol. 9, 4996-5007 (2009).

Chapter 2

2. Understanding the Degradability of sp2 Hybridised Carbon Nanomaterials for Their Applications in Biomedicine

Leon Newman, Cyrill Bussy* & Kostas Kostarelos*

This article has not been published

Affiliations: Nanomedicine Laboratory, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, M13 9PT, UK.

2.1. Statement.

This review article has not been published, but has been written to suit the specifications of the Chemical Society Reviews. Leon Newman wrote this review article, analysed the previous literature studies and generated the figures unless stated otherwise. Dr Cyrill Bussy and Prof Kostas Kostarelos critically reviewed the document and provided constructive feedback.

Key word: sp2 hybridised carbon nanomaterials, graphene, graphene oxide, carbon nanotubes, carbon nanohorns, fullerenes, degradation, enzymes, in vivo, in vitro

______

* Correspondence to either: [email protected]

[email protected]

2.2. Abstract.

sp2 hybridised carbon nanomaterials such as graphene, carbon nanotubes, carbon nanohorns and fullerenes are being intensively investigated in biomedicine and other areas of science and technology. Recent studies have demonstrated their promising potential and applicability to areas such as drug delivery, tissue engineering, biomedical imaging as well as electronics, water purification and paints. Successful application of these materials is predicated upon an in-depth comprehension of the fate of these unique materials within their application specific environments, for example the biological milieu in the case of biomedicine. One aspect of their fate that must be understood is their degradability and more importantly from an engineering perspective, what factors determine the kinetics of this process. An understanding of this is crucial and will contribute towards the safe, efficacious and quality design and implementation of these novel materials. In this review, we explore the degradability of various sp2 hybridised carbon nanomaterials mediated by various agents of increasing complexity (chemicals, enzymes, cells and finally whole organisms) and highlight various factors that can determine and influence the degradation kinetics. Using case studies, we then discuss the necessity of such knowledge by discussing the way in which the degradability of sp2 hybridised carbon nanomaterials must be carefully considered and tuned to allow for their successful application in various areas of biomedicine.

2.3. Introduction.

The carbon atom is distinct; it has 4 valence electrons that allow it to readily interact with other atoms via covalent interactions. The uniqueness of carbon however resides in its ability to form covalent bonds with atoms of its own element in different hybridisation states.1,2 This has facilitated the formation of a class of kinetically stable allotropes2, including the well-known hard, clear and electrically insulating diamond, that is composed of sp3 hybridised carbon.3 Another carbon allotrope found in nature is graphite, which can be described as a brittle, opaque and conductive material, composed of stacked arrays of 2D sheets based on sp2 hybridised carbon known as graphene.4 In recent years, nano- dimensional allotropic forms of sp2 hybridised carbon have also come to light. They include fullerenes5, carbon nanotubes (CNT)6, 7 carbon nanohorns (CNH)8, and graphene9, which can all be naturally created as a result of combustion, but are more consistently produced using chemical engineering processes.10-12 This superfamily of carbon nanomaterials (CNMs) has been attributed with outstanding properties including exceptional electronic, mechanical, chemical, optical and thermal properties, which can be ascribed to the combination of their crystalline sp2 hybridised carbon structure and nanoscale dimensionality. These unique properties have rapidly gained scientists’ and industries’ interest. Through scientific research within the fields of physics, chemistry, material science, as well as biomedicine, it is hoped that such materials will help develop a new range of ground breaking technologies13-15 alongside their current commercial applications.14, 16

As a result of this rising interest to use these materials, there is a growing demand for a detailed understanding of their fate and stability upon interaction with living organisms. This is important both for helping to incorporate CNMs into products in a suitable and safe manner, and to further understand the safety implications that their widespread use may have on humans and also the wider environment.17 One aspect of the fate that is crucial to understand is their biodegradability, or the susceptibility of CNMs to be broken down to smaller components with the concurrent loss of their properties via biologically mediated processes. Studies have now begun to suggest that, contrary to past ideas which negated such a concept, biodegradation is possible under specific conditions.18, 19 In this review, we aim to discuss the different systems that have been reported to degrade CNMs and identify in each case the factors that can influence the kinetics of CNMs biodegradation. We briefly mention the toxicological implications of the resultant CNM degradation by-products.20-22 Finally, we discuss degradation in the context of three different biomedical applications of

CNMs in order to highlight the value and need to consider degradation in these applications. We further discuss based on the previous section how to engineer CNMs to allow their degradation kinetics to fit the specific requirements of particular biomedical applications.

2.3.1. sp2 hybridised CNMs.

In order to fully appreciate the biodegradability of CNMs, we first begin with a brief overview of the CNM structures that will be considered. This includes fullerenes, CNTs, carbon nanohorns (CNHs) and graphene, as well as their functional derivatives.

2.3.2. Fullerenes.

Fullerenes were the first of the sp2 hybridised allotropes of carbon to be identified.5 They were initially characterised via mass spectrometry and their structures determined with the help of Buckminster Fuller’s architectural models. They were described as a ‘remarkably stable structure consisting of 60 carbon atoms’ organised into a highly curved truncated 5 icosahedron. In order to suit the truncated icosahedron geometry, the C60 structure is composed of 12 pentagonal and 20 hexagonal rings, where each pentagonal ring is positioned at the vertices of an icosahedron.5 An illustration of this structure is shown in Figure 2.1. The material was first thought of as a highly aromatic species; however, subsequent studies have shown that this is not the case; instead fullerenes behave as highly strained electron deficient alkenes.23, 24 Moreover fullerenes display poor solubility in many polar solvents, which in terms of their applicability to various areas of technology is a major caveat. Nonetheless due to their electrophilic character, nucleophilic addition reactions are feasible. Accordingly much effort has been devoted to covalently functionalise these nanostructures enabling their facile application in a variety of fields such as biomedicine.24, 25 During these functionalisation reactions, the high strain associated with the curved sp2 bonded material means that the hybridisation state of the carbon will readily convert from sp2 3 (planar) to sp (tetrahedral), thereby relieving strain. Since the initial discovery of the C60 fullerenes, many more fullerene structures have been identified that are composed of varying numbers of carbon atoms.26 Moreover the characterisation of these structures has advanced, with various studies reporting electron microscopy based visualisation of individual fullerenes, as well as their formation under different conditions including arc- discharge and laser ablation, in addition to under electron beams.12, 27

Figure 2.1 An illustration of the typical structure of a truncated icosahedron C60 fullerene.

2.3.3. Carbon nanotubes.

Carbon nanotubes (CNTs)6, 7, like fullerenes are another class of CNMs. CNTs were initially identified as high aspect ratio multiwalled tube-like structures, termed multi-walled CNTs (MWNT), composed of at least 2 concentric cylindrically arranged 2D sheets (or walls) of graphene. Within these cylindrical walls, the sp2 carbon atoms are helically arranged as a highly aromatic hexagonal honeycomb lattice about the centre of the tube. The graphitic structure of the material was confirmed with electron diffraction and high resolution transmission electron microscopy. In a pristine MWNT, each wall is separated by a distance of 0.34 nm, consistent with the interlayer distance of non-defected graphite.6 The hollow centre of this tube has been measured to vary between 0.4 nm to a few nanometres and continues for the length of the tube (1 μm to a few centimetres)6 (Figure 2.2). In pristine MWNTs, the edges of the tubes are usually capped by dome-shaped half fullerenes structures which incorporate pentagonal defects allowing the necessary curvature. The MWNT structures can be further separated into two categories - the Russian doll model or the parchment model. In the Russian doll model, there are two or more concentric cylindrical walls, where the outer walls have greater diameters than the preceding inner ones. In the parchment model, the nanotube can be described as a rolled 2D graphene sheet28; in these cases the structures are commonly termed ‘nanoscrolls’. MWNTs have been attributed with various outstanding properties such as high tensile strength, enhanced electron mobility, and heat conductivity.28

A B

Figure 2.2 An illustration of the structure of MWNTs.

(A) A typical model of a MWNT (capped ends not shown) B) HR-TEM micrographs of 5-, 2- and 7 walled MWNTs shown from left to right. Reprinted by permission from Macmillan Publishers Ltd: [Nature] (S. Iijima, Nature, 1991, 354, 56-58), copyright (1991).

In addition to MWNTs, single-walled carbon nanotubes (SWNT) have also been identified.7 SWNTs are composed of a single atom thick, seamless graphene-based wall with diameters ranging from 0.4 to 2 - 3 nm and a length within the sub-micrometre to micrometre range.7 Depending on the orientation of the graphene walls, one can distinguish three forms of SWNTs namely arm chair, chiral and zig zag as illustrated in Figure 2.3. The structure of the SWNT can be described by a pair of indices (n,m) that can be used to identify the different specific forms of SWNT. In general, where m = 0, the nanotubes are termed zig zag and where m = n the nanotubes are described as arm chair; other states are termed chiral.28, 29 Each of these forms has a different orientation due to the ways in which the graphene wall is orientated. The fundamental properties of each nanotube such as the electrical and optical properties will vary accordingly.28 When comparing MWNTs and SWNTs, the overall attributes are similar; however the resistance to twisting and tensile strength is greater for MWNT than for SWNT. In terms of chemical reactivity, the stacked walls of MWNTs can provide some shielding of their inner walls from chemical modification by reactive substances. This is not possible for SWNTs, meaning that theoretically a single SWNT would be more susceptible to rapid degradation than a single MWNT.28

Zig Zag

Chiral

Arm Chair

Figure 2.3 An illustration of the different forms of SWNT that can be derived from a 2D sheet of graphene (zig zag, arm chair and chiral).

2.3.4. Carbon nanohorns.

Other less well known tubular CNMs have also been identified, including the single-walled carbon nanohorns (CNH) that were first observed in 1999.8 These can be described as cone-shaped carbon nanostructures with typical diameters of 2 - 5 nm and lengths of 30 - 50 nm. The end of the cone shaped structure of minimum diameter can be viewed as a capped

30 horn-shaped tip.8, 30 This cap is composed of a half fullerene-like structure and similar to pristine CNTs, incorporates pentagonal rings allowing the necessary curvature. In the as produced state, CNH exist in 80 – 100 nm dahlia-like aggregates composed of approximately 2000 individual CNH8, 31 as shown in Figure 2.4. Theories explaining the 32 formation of the dahlia-like form are numerous . For example, they may form via C2 and C3 fragment gas phase recombination, similar to what was described for fullerene synthesis.33 Harris et al. proposed that random gas phase collisions of aerosolised fragments of graphite would not be sufficient to allow the pentagonal rings to optimally position, allowing the complete closure of the structures into the cage like truncated icosahedron fullerenes34. This was thought to be due to the relatively low temperatures and short annealing times that occur during CNH synthesis. Instead they form CNHs and cluster as dahlia-like structures.34 However considering the somewhat haphazard nature of these processes, it is unlikely to explain the relatively narrow size distributions of both dahlia-like structures and their component nanohorns. A more rigorous proposal was presented by Yudasaka and co- workers, who explained that dahlia-like structures may develop following the generation of liquid phase carbon droplets that result from the high temperatures generated during laser ablation of graphite.35 The liquid phase carbon could then transform into single 2D sheets of graphene, which close to form nanohorns, associated in dahlia-like structures, under specific conditions of temperature and carrier gas pressure. This theory has been supported by tight binding molecular dynamic simulations36 as well as experimental studies, in which the carrier gas pressure was altered. This in turn resulted in the formation of crumpled graphite platelets of comparable dimensions instead of true dahlia-like structures.37

A B

Figure 2.4 An illustration of the structures of CNHs and dahlia-like superstructures.

A) A typical mode of dahlia-like super structure of individual CNHs (top) and individual CNHs (bottom) B) HR- TEM micrographs of dahlia like super-structures and a magnification of the constituent CNH showing their horn shaped tips. Reprinted from Nano-aggregates of single-walled graphitic carbon nano-horns, 309, S. Iijima,M. Yudasaka,R. Yamada,S. Bandow,K. Suenaga,F. Kokai,K. Takahashi, Pages No. 165 - 170, with permission from Elsevier.

2.3.5. Graphene.

The newest addition to the CNM family is graphene.9 This material can be seen as the archetypal member of the other described CNMs13, 38 as has been expounded in Figure 2.5. It exists as a single atom thick sheet of sp2 hybridised carbon atoms. Within this structure, the carbon atoms exist as an aromatic hexagonal honeycomb lattice. Based on various theoretical models39, 40, the 2D material was previously predicted not to exist due to thermal instabilities, which could force the sheets to reconfigure to form other carbon allotropes. However, following a simple experiment involving graphite and sticky tape, graphene was isolated and electronically characterised.9, 13 Graphene, being the underlying template material of other CNMs, is attributed with properties that are similar and enhanced compared to some of other sp2 carbon nano-allotropes. For example, graphene has been shown to possess various remarkable optical, mechanical, chemical, electronic and thermal

32 properties.13 Moreover the material displays some characters which are highly unique such as quantum Hall effect that can be observed at room temperature.41

Graphene sheet

Graphene CNT Fullerene nanosheets

Figure 2.5 An illustration of graphene as the archetypal member of the family of CNMs.

Reprinted by permission from Macmillan Publishers Ltd: [Nature Materials] (A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183-191), copyright (2007).

2.3.6. Functionalised derivatives.

In addition to the pristine CNMs, a multitude of functionalised (covalent or non-covalent) derivatives have also been produced and investigated. These include but are not limited to hydrides, hydroxides, halides, oxides, alkyl, aryl as well as aromatic fullerenes derivatives24, oxidised or aminated CNTs42, oxidised CNHs43, graphene oxide (GO)44, aminated graphene45 as well as PEGylated graphene derivatives.46 In these structures, the added chemical functionalities provide the nanomaterials with additional properties that allow their usage in different fields. For instance, the polar oxygen containing functionalities of GO or

33 oxidised CNTs confer increased water dispersibility to these CNMs. This allows more facile application of the material in biomedicine, for example as intravenous drug delivery nanovectors, where good dispersibility is pertinent to avoid agglomeration in aqueous environments such as the blood stream.47

The most intensively researched functionalised derivative for graphene is GO. GO can be described as graphene sheets decorated with various oxygen functional groups including epoxide, hydroxyl, carbonyl and carboxyl groups48 as shown in Figure 2.6. There is a lot of interest into the development of GO for various applications because of its large surface area, its dispersibility in water, its unique optical properties, as well as its ease of functionalisation.48, 49 Surprisingly, GO has been known for decades and was first identified by Brodie in 1859, when he was exploring the physicochemical properties of graphite50. Since then, other iterative methods have been proposed. But three methods are considered as seminal for the others: the Brodie method50, the Staudenmaier method51 and the Hummers and Offeman method.52

Figure 2.6 The structure of GO.

Following chemically induced oxidation of graphite, GO can be obtained which is functionalised with various oxygen functional groups such as carboxylic acid, alcohol and epoxide groups. The structure contains a mixture of both sp2 and sp3 carbons as a result of the chemical treatment. The added hydrophilic groups allow the molecule to participate to some extent in the hydrogen bonding network of water resulting in greater water dispersibility, relative to graphene, permitting application in various areas of biomedicine.

As is true for the preparation of most functionalised derivatives of CNMs, the physicochemical features of GO can differ between batches due to differences in reaction conditions, purification processes as well as the starting graphite material.48, 53 Ultimately, this means that the term GO can be considered an umbrella term for a variety of related forms of oxidised graphene, and should only be used in conjunction with a detailed physicochemical characterisation of the materials used to avoid ambiguities. Efforts have been made to stress the need for using a clear nomenclature system, in view of biological based studies which are sensitive to changes in materials’ physicochemical properties.38, 54, 55 In particular the ‘classification framework for graphene based materials’ was proposed by the European Graphene Flagship Future and Emerging Technologies Project.56 This framework considers 3 features – the thickness (in terms of number of layers), the lateral dimensions, and the extent of oxidation as defined by the carbon to oxygen ratio – to be the key parameters to define graphene-based materials, including GO. These 3 parameters were chosen as they represent some of the main factors that can vary between different GO materials and between batches, as well as between individual GO sheets within the same batch of material.38, 56 These parameters are also considered because they have a major impact on the materials’ specific biological properties, particularly their biodistribution57, 58 and biodegradability59, which in turn can change their suitability to specific biomedical applications.

2.4. Degradation and the use of CNMs in biomedicine.

One area in which CNMs and their functionalised derivatives have received great interest is biomedicine.15, 60 The outstanding properties of these materials can be leveraged to enhance several biomedical areas including drug delivery, regenerative medicine, as well as diagnostics and monitoring.15, 38, 60, 61 If CNMs are to be developed for biomedical purpose, the revelations of their safety profile will be an essential requirement.62 Not only will various toxicological endpoints have to be checked, but the biodurability of specific CNMs in living organisms or the biological milieu (i.e. biodegradability) will need to be assessed.

Prior to 2008, the concept of CNM degradation under physiological or even environmental conditions was repudiated 63; this was on account of the associated unreactive and highly stable sp2 graphitic backbone. Within the last decade however, there has been a growing number of reports evidencing that these unique materials can undergo significant structural alterations and degradation to simpler components via biological processes (biodegradation) both within the

35 context of biomedicine as well as the environment.18, 63, 64 With regards to the use of these materials in biomedicine, an understanding of whether the material is biodegradable or not, although informative, is not sufficient for the development of safe, quality and efficacious products. It is also essential to understand what factors can influence the kinetics associated with the degradability, and to determine whether these degradation kinetics are desirable with regards to a specific targeted application.

In the following section, we review information from the available literature studies to illustrate the wide range of systems of increasing complexity that have been shown to degrade CNMs, we pay particular attention to the parameters can influence the kinetics of the degradative process at each level of complexity.

2.5. Chemically mediated degradation.

Chemical methods have been used to functionalise and manipulate the physicochemical properties of CNMs for decades. Strong oxidising agents such as hydrogen peroxide, hypochlorite, nitric acid, or sulphuric acid have been used to produce oxidised derivatives of fullerenes65, 66, CNTs42, CNHs43 and graphene.53 Various studies have looked systematically at the changes induced by different chemical treatments to the physicochemical properties of CNMs. Key examples of this is provided by Brodie50, Staudenmaier51, Offeman and Hummers.52 These chemists used harsh oxidants to change the physical properties of graphite to form graphite oxide that is attributed with high defects and hydrophilicity, as well as an acidic character unlike the parent graphite material. Moreover, research into the behaviours of these materials under different conditions has resulted in more efficient and streamlined protocols for the production of single- to few-layer GO flakes.44, 53, 67, 68

In addition to this, detailed chemical analysis has shown that oxidation procedures can fracture, exfoliate and decorate the nanomaterials with various oxygen functionalities such as hydroxyl, epoxide, carbonyl, carboxylic acid as well as lactone groups on the edges of the GO sheet.48 Similar studies have also been completed on CNTs, CNHs and fullerenes, and have confirmed that oxidation leads to defects in the sp2 carbon structure of the CNMs. For example, MWNT have been exposed to oxidising conditions evoked during photo-Fenton and normal Fenton chemistry, during which the resulting oxidation of MWNT was characterised over time.69 Some strategies employing harsh treatments including argon

70 71 72 73 plasma , KMnO4 , NaClO , HNO4 and H2SO4 can lead to the unzipping of CNTs to produce narrow single-layer graphene-based sheets, also named graphene nanoribbons. Light-assisted oxidation with hydrogen peroxide has also been shown to promote oxygen- rich functionalisation and the formation of holes in the walls and tips of CNHs.43 The effects of oxidising chemicals on fullerenes has been extensively studied to overcome their major caveat – their limited solubility in aqueous solvents.24

Overall, it is evident that oxidising chemicals and harsh conditions can mediate drastic changes to the CNM crystalline structure (i.e. structural defects to the C=C arrangements). These studies constitute the foundations that can provide clues concerning the sorts of chemical species and conditions that could lead to the degradation of CNMs in biological milieus. Indeed, chemical functionalisation via oxidation is not far from oxidative conditions and biodegradation processes observed in living organisms. For example, it has been shown that when graphene (when bound to a solid support) is incubated in environmentally and physiologically relevant concentrations of H2O2 for 25 h, the pristine sheet structure undergoes gross structural changes which includes the formation of holes and amorphous carbon structures.74 By the end of the study, the spectroscopic signature of graphene however could still be identified, indicating that degradation was not fully complete. Nevertheless, this study shows that the simple application of environmental and biologically 74 relevant concentrations of H2O2 have the capability to partially degrade graphene. Such changes can lead to alterations in the materials’ properties, for example its mechanical properties.75 This becomes highly relevant if the materials’ mechanical properties are essential for successful application.

This study’s findings were furthered when 14C-labelled graphene was demonstrated to undergo complete degradation in high concentrations of H2O2, beyond those encountered naturally, in the presence of an iron (Fe) catalyst.76 Fe is expected to enhance the 77 generation of reactive oxygen species (ROS) from H2O2 via the Fenton reaction . On account of the intrinsic nature of the 14C labelling, the authors76 were able to quantitatively 14 determine the kinetics of the degradation by following the release of C labelled CO2. At high concentrations of H2O2, the structural changes of graphene and decomposition into CO2 occurred rapidly within 3 days, emphasising the role that relevant catalysts can play in enhancing degradation kinetics. However, under environmentally relevant conditions, the process occurred at substantially lower rates than what was reported previously.74 Discrepancies between the two studies are potentially a result of differences in the materials’

37 physicochemical characteristics. Moreover, in the study by Xing et al. the graphene sheet was fixed on to a silicon surface74, while in the case of Feng et al. this was not the case, which increases the availability of graphene to undergo oxidative degradation.76

Nevertheless, both studies have shown that H2O2 is capable of inducing structural defects in graphene, and this oxidative degradation can be enhanced in the presence of Fenton catalysts.

Using a similar strategy, Bai et al. explored the propensity of high concentration of H2O2, to degrade a derivative of GO in the presence of a Fe catalyst under UV-irradiation, as part of the photo-Fenton reaction.78 The aim of the study was to elucidate the mechanism by which degradation of GO sheets would occur. The results indicated that under the prevailing conditions, the GO flakes sustained severe defects including the complete disintegration of the planar GO sheet-like morphology. The authors reported that as part of this process, short-lived low molecular weight oxidised poly aromatic hydrocarbon (O-PAH) by-products were formed temporally but were rapidly degraded to CO2 by day 3. Concomitantly by day 3, the solution was increasingly dominated by highly fluorescent graphene quantum dots, thought to form due to the oxidative rupturing of epoxide groups into carbonyl groups which 2 ultimately led to the formation of nanometre-size aromatic sp domains surrounded by oxygen functionalised edges (Figure 2.7). Recently, this degradation method of GO was optimised towards the development of an economical method to degrade the material in wastewater, under the assumption that such a method will be soon required considering the increase in GO production and research.79 These experiments highlight the impact that the concentration of reactants, the presence of relevant catalysts and radiation can have on enhancing degradation kinetics. These factors have a role in helping to destabilize the peroxide in order to generate more ROS, thereby increasing the kinetics of degradation. Such a strategy could be useful with regards to GO, for which concerns have been raised regarding its high water dispersibility and therefore its potential to be transported in groundwater and surface water as a pollutant.80

Figure 2.7 A schematic diagram demonstrating the facile degradability of GO via the photo- Fenton reaction.

(A)_During the reaction, low molecular weight oxidized polycyclic aromatic hydrocarbon (o-PAH) temporary intermediates are produced. Towards the latter stages graphene quantum dots (GQDs) accumulate with time. (B) Atomic force microscopy images following the fate of surface bound GO during the photo-Fenton at time points 0, 18, 36, and 54 h. All images were obtained via tapping mode, scale bars represent 500 nm. Adapted with permission from H. Bai, W. Jiang, G. P. Kotchey, W. A. Saidi, B. J. Bythell, J. M. Jarvis, A. G. Marshall, R. A. Robinson and A. Star, J. Phys. Chem. C Nanomater. Interfaces, 2014, 118, 10519-10529. Copyright (2014) American Chemical Society.

This photo-oxidative degradation has been demonstrated in more environmentally relevant scenarios. Where instead of a UV light source, sunlight was used to help destabilise the 81 H2O2 that in turn degraded the GO sheets. Under this less extreme oxidative influence, GO was more gradually oxidised to CO2 via various temporary intermediate carboxylated by- products, that had lower total oxygen contents than the parent GO. It is plausible to suggest that these temporary reduced structures could be related to the graphene quantum dots observed by Bai et al.78 The authors81 proposed the presence of such temporary structures could result from the parallel sunlight-induced GO photo-reduction, which they had demonstrated previously.82 Collectively, these studies show that the different conditions,

39 including the presence of light or catalysts, in which H2O2 is used to degrade graphene- based materials determines its success and kinetics. Therefore, the nature and concentration of the degradation medium must be taken into account when considering CNM degradation. For instance, on its own, the ability of H2O2 to break down GO appears to be somewhat limited, particularly at lower concentrations. However in the presence of metallic

Fenton catalysts or light, the H2O2 mediated GO degradation becomes more efficient. In addition to this, the dynamics of degradation can change, with the production of various temporary by-products under different circumstances.

In addition to investigations employing H2O2, several studies exploring the degradation of CNMs in living systems have used NaClO as positive controls. This is on account of its high oxidising capacity and relevance to biological environments as its production can be catalysed by various members of the peroxidase enzyme family. For example, NaClO has been shown to be able to efficiently degrade oxidised fullerenes.83 Several studies have also demonstrated the NaClO mediated degradation of CNTs.84, 85 Curiously, the influence of CNM surface protein coatings has been shown to effect to the kinetics of NaClO mediated CNM degradation.86 The degradation kinetics were reduced when the materials were coated with proteins. It is envisioned that the protein coating could serve as a protective barrier from the oxidising agents, thereby reducing degradation kinetics (Figure 2.8).86 This finding is highly relevant to biological contexts, such as in the blood plasma and in cells, where proteins are prevalent and can spontaneously interact with CNM surfaces. It can be used to inform further studies investigating the use of NaClO or other isolated oxidising chemicals for the removal of CNM waste that is likely to be contaminated with proteinaceous or other organic matter waste.

A B E

C D

Figure 2.8 Binding of human serum albumin (HSA) to SWNTs reduces the kinetics of NaClO- mediated degradation in vitro.

(A) Photograph of the various nanotube suspensions followings 2 days incubation with NaClO. (B) Vis-NIR spectra of different nanotube suspensions after 2 days, demonstrating the loss of SWNT M1 and S2 bands during the NaClO-mediated degradation. (C) Raman spectra of different SWNT suspensions after 2 days with an insert highlighting the radial breathing mode (RBM) region of the respective Raman spectra. (D) ID/ IG intensity ratios obtained using Raman spectroscopy on SWNTs at different indicated time points, showing a trend of increasing defects on the walls of SWNTs (*p < 0.05, groups versus untreated-nanotubes). (E) SEM analyses tracking the degradation of nanotubes after 3 days incubation with NaClO. Adapted with permission from N. Lu, J. Li, R. Tian and Y.-Y. Peng, Chem. Res. Toxicol., 2014, 27, 1070-1077 Copyright (2014) American Chemical Society.

In addition to isolated chemicals, more biologically relevant artificial liquid mixtures, representing a higher level of complexity, have also been tested for their ability to degrade CNMs. Phagolysosomal simulant fluid (PSF), used to emulate the chemical contents of phagosomes (intracellular lipid-bound vesicles formed after phagocytic cells internalise particulates via phagocytosis), has been shown to degrade SWNTs in a functionalisation dependent manner.87 Carboxylated SWNTs underwent degradation in PSF over 90 days whereas pristine, ozone-treated, and aryl-sulfonated SWNTs failed to degrade under the same conditions.87 It was thought that carboxylation via harsh acidic treatments induces more severe oxidative defects than ozonolysis or aryl-sulfonation, which in turn increases the reactivity of the carboxylated SWNTs to PSF and therefore their degradability.

The degradation of functionalised SWNTs in comparison to MWNTs in PSF has also been explored.88 In this study, the authors compared the degradation of oxidised SWNTs and oxidised MWNTs. The results demonstrated that oxidised MWNTs were degraded with far slower kinetics than oxidised SWNTs. After 60 days, while all SWNTs were degraded, tubular structures of oxidised MWNTs could still be identified. This is likely due to the protective effects of the multiple concentric walls present in MWNTs, not present in SWNTs. In strong contrast, when pristine SWNTs and MWNTs were incubated in lysosome simulated biological fluid (i.e. Gamble’s solution) little or no change was seen for most of the samples.89 This was most likely due to the lack of reactive sites on the surface of pristine CNTs with which the oxidising chemicals could interact and thereby catalyse degradation. These studies underline the potent role that surface defects, chemistry and structural differences have in driving the kinetics of CNM degradation mediated by isolated reactants.

2.5.1. Overview of chemical mediated CNM degradation and the lessons learnt.

Overall, chemically mediated degradation of CNMs can occur through the action of various highly reactive species such as highly oxidising chemicals. Moreover, the greater the reactivity of the species, the more rapid the degradation kinetics will be. Higher amounts of reactive species can be obtained for example, by the use of catalysts or light to destabilise the reactant molecules to produce free radicals. A key example of this is the photo-Fenton reaction. In terms of the material characteristics, thin or single-walled, highly oxidised or well dispersed structures, which are not contaminated with proteins or other macromolecules, undergo a more rapid degradation than others. A summary of all the factors shown to effect degradation kinetics has been summarised in Figure 2.9. An exhaustive table (Table 2.1) has also been given that lists all the published studies that have specifically focused on exploring CNMs mediated by isolated chemicals.

Structural Protein Defects coating

Chemical mediated Specific CNM biodegradation catalysts and Kinetics radiation

Few layers (or Low walls) concentrations of reactants

Figure 2.9 A summary of the factors that can influence the kinetics of chemically mediated CNM degradation.

Where red arrows indicate acceleratory influences on the kinetics (indicated by the + symbol) while green arrows indicate retarding factors on the kinetics (indicated by the – symbol).

Table 2. 1 A list of the literature studies that have specifically focused on the chemically mediated degradation of CNMs.

Chemicals

Type Modification Model Summary of Findings Reference

87 SWNT Ox PSF  Only oxidised SWNT were observed to degrade.  Low pH is not a necessity for degradation.

89 SWNT Pristine Gamble’s solution  Long MWNT lost 30% of their original mass within the first 3 weeks of incubation, but show no further modification thereafter. MWNT  After incubation of MWNT for 10 weeks the proportion of long fibres had decreased, but tubular structure could still be identified.  SWNT underwent significant degradation.

Graphene Pristine H2O2  Surface bound graphene degraded in physiologically and environmentally relevant concentrations of 74

H2O2.

78 Graphene Ox UV irradiation, H2O2 and  GO sheets underwent severe structural deformations. Fe  After 1 day incubation of GO under the reaction conditions, intermediate oxidation products (with MW 150−1000 Da) were generated.  Upon longer reaction times (i.e., days 2 and 3), the 150−1000 Da molecular products disappeared and the system was dominated with graphene quantum dots.

14 Graphene C enriched H2O2  After 5 days incubation of graphene with high concentrations of H2O2, there was significant 76

mineralisation of graphene to CO2.

 Substantially lower rates of CO2 evolution from graphene were detected under environmentally relevant concentrations.  Significant physicochemical changes in the graphene sheets occurred during the reaction.  Multiple soluble intermediate degradation by-products were identified.  Substantial bioaccumulation of non-degraded graphene sheets were measured in Daphnia magna after 24 h, while accumulation of the degradation intermediates was nearly two orders of magnitude less.

82 Graphene Ox Simulated solar light  GO was shown to photo-disproportionate to CO2, small reduced fragments and low molecular-weight species.  The concentration of dissolved oxygen effected the rate of mineralisation of low molecular weight 44

species to CO2 but not the rate at which initial photo-disproportionation of GO occurs.  rGO species by-products of the photo reaction persisted for up to two months.

81 Graphene Ox H2O2 in the presence of  GO was extensively photo-decomposed to CO2, in the presence of H2O2 when irradiated for 48 h. direct and indirect  Reaction with •OH causes increases in the concentrations of carboxylic acid groups of photo-reacted simulated sunlight GO and low molecular- weight (LMW) species as part of the intermediate photoproducts.  Intermediate by-products interact less with cell membrane models than the original GO materials.

79 Graphene Ox UV irradiation, H2O2 and  GO in wastewater can be most efficiently degraded under the following conditions: pH value of 3,

Fe concentration of H2O2 and FeCl3 are 35 mM and 5 ppm, respectively.

In vitro chemically mediated degradation studies. PSF = Phagolysosomal simulant fluid, Ox= oxidised or carboxylated.

2.6. Enzyme catalysed biodegradation.

In addition to their generation under artificial conditions, oxidising chemicals can be generated via the action of enzymes (biological catalysts) derived via natural or synthetic means. Accordingly, studies have been performed to evaluate the susceptibility of fullerenes, CNTs, CNHs as well as graphene-based structures to undergo enzyme-catalysed biodegradation.

The most highly investigated CNM with regards to enzyme mediated biodegradation is currently the CNTs. This could be due to the pathogenicity to the pulmonary system of some CNTs that has been compared to asbestos fibres.90-92 One of the main reasons for asbestos toxicity is due to their bio-persistence in the lungs, and accordingly this has raised similar questions regarding CNTs’ ability to resist biodegradation processes in living systems.

2.6.1. The peroxidase enzyme superfamily.

The peroxidase family of enzymes are currently the most intensively studied group of enzymes regarding the biodegradation of CNMs. These enzymes, known as peroxidases, are characterised by the use of various peroxides (R-OOH), such as hydrogen peroxide

(H2O2) as electron acceptors to catalyse a plethora of oxidative reactions via the production of various reactive intermediate species. The enzymatic activity is made possible by their metallic catalytic core. The core is usually based on Fe, but it can be based on other transition metals too. Members of this enzyme superfamily are widespread in all living organisms.93 For example in mammals, they are involved in various biological functions including immune and hormonal regulation as well as growth. The enzymes also occur in the plant kingdom where they are implicated in auxin regulation, immune functions, as well as growth and development.94 The ability of these enzymes to oxidise biological molecules, such as aromatic species and alkene bearing molecules including proteins, lipids and carbohydrates, makes them promising candidates for the biodegradation of CNMs.

2.6.2. Horse radish peroxidase.

The horse radish peroxidase (HRP) enzyme, which contains an Fe catalytic core, was the first enzyme piloted to prove the enzyme-catalysed biodegradability of CNMs. Allen et al. conducted the initial study that described how carboxylated SWNTs can be biodegraded

46 over 12 weeks at 4°C by the oxidative action of HRP in the presence of low concentrations 95 (800 μM) of H2O2. The high redox potential of HRP Compound I/ Compound II cycle in the presence of H2O2 was responsible for the efficient catalysis of the oxidative biodegradation.

Low concentrations of either Fe or H2O2 applied alone were insufficient to produce the same biodegradation kinetics, indicative of the importance of the enzyme protein structure, particularly the active site. The authors completed a follow-up study in which the same experiment was completed over 10 days at 25°C.96 Biodegradation occurred with more rapid kinetics at the higher temperature. This is expected as per the Q-10 effect, which describes the relationship between enzyme activity and increases in temperature.97, 98 Moreover, it was shown that the carboxylated SWNTs underwent mineralisation to CO2 via various temporary intermediate by-products, similar to what was previously shown for photo-Fenton reaction mediated biodegradation of GO.78 The authors also tested the biodegradability of non- functionalised pristine SWNTs. These pristine SWNTs did not biodegrade under the same conditions.96 Computer modelling demonstrated that on account of the enzyme conformation, and the surface properties of the pristine SWNTs, there was an increased distance between the enzyme’s Fe catalytic core and the pristine SWNT surface that hindered biodegradation. This was further supported by the fact that both carboxylated and pristine nanotubes could be degraded by H2O2 in the presence of Fe or hemin (an Fe containing porphyrin) alone, albeit at a lower rate, where protein conformation is not relevant.96 Taken together, these studies emphasise that both temperature and surface oxidation have important influences on HRP mediated biodegradation kinetics.

The same team of researchers investigated the application of HRP to biodegrade MWNTs with 4 different surface chemistries, namely pristine, highly carboxylated, intermediately carboxylated or nitrogen doped MWNTs.99 Biodegradation of the MWNTs was monitored, under the same conditions used previously96, but this time over 80 days. This was much longer than that needed to biodegrade SWNTs (i.e. 10 days)96 and was likely due to the protective effect of the multiple concentric walls, and also the more pristine character of the inner walls.88 For the carboxylated nanotubes, the greater the degree of carboxylation, the greater the rate and extent of HRP mediated MWNT biodegradation. This was to the extent that the biodegradative changes in the pristine sample were minimal. The study emphasises how surface properties of MWNTs can play a major role in determining biodegradation kinetics. Nitrogen-doped MWNTs underwent a more rapid biodegradation, with the complete disappearance of the tubular structure by day 50.99 It is thought that since nitrogen has one more electron than carbon, the atom is incompatible with the typical seamless hexagonal graphitic structure of MWNTs. In order to allow for its energetically favourable presence

47 within the MWNTs, the nitrogen-doped MWNTs formed as an intrinsically defected tubular structure that was compartmentalised into stacked cup-shaped sections (i.e. bamboo-like structures), where nitrogen only occurs at the edges of these cups. Therefore due to the multiplicity of defects and dangling bonds, the enzymes could interact more easily with the defected nitrogen doped MWNTs, thereby catalysing their biodegradation more efficiently compared to the other materials. This is a key example of how engineering of CNTs can be implemented in order to promote biodegradation.99

Increasing structural defects are however not the only means by which CNMs can be engineered to biodegrade with more rapid kinetics. We recently demonstrated that the attachment of functional enzymatic substrates onto oxidised MWNTs could promote the biodegradation of the tubular sp2 hybridised carbon nanostructures by HRP.100 We found that coumarin and catechol substrates were efficient biodegradation enhancers of oxidised MWNTs by HRP. On the other hand, purine substrates could delay the biodegradability of MWNTs by xanthine oxidase (XO) (Figure 2.10). Such studies are crucial in increasing our understanding of how to tailor the biodegradability of CNMs to suit specific applications.

A B i i

ii iii ii iii

C i D i

iii iv

ii iii ii iii

Figure 2.10 TEM images demonstrating how the enzymatic biodegradation of MWNTs can be altered via the conjugation of specific substrates.

Different MWNTs were functionalised to substrates derived from (A) 3-azido-7-hydroxycoumarin (substrate for HRP), (B) 4-methyl-7-(azidopropyl) coumarin (substrate for HRP), (C) 3,4- dihydroxybenzoic acid (substrate for HRP) and (D) 9-(3-azidopropyl)-purine (substrate for XO). For each functionalisation strategy, i) a schematic of the resultant functionalised CNT is given together with the morphology of the different functionalised substrates that was interrogated by TEM at ii) 0 and iii) 20 h post incubation with their respective enzyme solutions. Adapted from Degradation-by- design: Surface modification with functional substrates that enhance the enzymatic degradation of carbon nanotubes, 72, Adukamparai Rajukrishnan Sureshbabu, Rajendra Kurapati, Julie Russier,Cécilia Ménard-Moyon, Isacco Bartolini, Moreno Meneghetti, Kostas Kostarelos,Alberto Bianco, Pages no. 20 – 28, with permission from Elsevier

The biodegradative capacity of HRP has also been investigated with respect to GO and 101 reduced GO (rGO). HRP could catalyse the biodegradation of GO to CO2 within 20 days.

49 rGO on the other hand, which contains fewer oxygen functionalities did not undergo biodegradation. This finding echoed the results of previous studies which explored the biodegradability of CNTs relative to the surface functionalisation and degree of oxidation.96 Through computer simulations, it was shown that when an HRP enzyme was docked with a GO sheet, two different potential sites on the enzyme could interact with the GO sheet (binding energies -24.8 and -22.4 kcal mol-1), while for rGO, there was only one (binding energy -26.7 kcal mol-1).101 This meant that on account of the differences in the material surface characteristics, there was a restriction in the HRP enzyme’s mobility when bound to rGO compared to GO, thereby hindering the enzymes movement which was required for its efficient catalysis of biodegradation. This hypothesis was further substantiated by electrophoresis based data. The study demonstrates that the rate of biodegradation can be altered by tuning the surface characteristics of GO sheets, in a similar manner to CNTs.

This concept has been further explored for the design of GO materials with tailored surface chemistries that remarkably would only biodegrade under specified physiological replicated conditions.102 In this study, when the intrinsically defected GO sheets were non-covalently coated with polyethylene glycol (PEG) or bovine serum albumin, their biodegradation kinetics were significantly attenuated compared to the uncoated GO sheets. It is possible that the coatings provided non-specific shielding of the surface of GO sheets from the enzymes as previously described in NaClO mediated CNM degradation86. In order to exploit this, GO sheets were functionalised with PEG via a disulphide bridge102. As expected, these constructs demonstrated poor biodegradation kinetics in the presence of just HRP and H2O2. However, upon cleavage of the disulphide bridge using a reducing agent (to mimic the reductive intracellular environment) the PEG groups were released and the GO sheets underwent facile HRP mediated biodegradation over 4 days.102 The results of this study have been presented in Figure 2.11. This study clearly emphasises that the surface chemistry of GO, like CNT, can be carefully engineered using various strategies to implement ‘degradation by design’ stratagems that can help tailor CNMs to specific applications, for example in biomedicine.

A B C

i ii iii iv

Figure 2.11 The condition specific biodegradation of disulphide-linked GO-SS-PEG.

(A) Structure of GO-SS-PEG, UV-vis-NIR spectra and inserted photos of 0.03 mg/mL GO-SS-PEG samples (B) without or (C) with (C) DTT pre-treatment, before and after HRP-induced biodegradation for 4 days. A. (D) TEM images of DTT-pre-treated GO-SS-PEG at day 0 (i and ii) and day 4 (iii and iv). Adapted from Y. Li, L. Feng, X. Shi, X. Wang, Y. Yang, K. Yang, T. Liu, G. Yang and Z. Liu, Small, 2013, 2014, 10, 1544-1554 with permission of John Wiley & Sons Inc.

2.6.3. Myeloperoxidase.

Following the studies using the originally piloted plant HRP, other more biomedically relevant peroxidase family members have also been explored for the purposes of CNMs biodegradation including myeloperoxidase (MPO). MPO is a cationic domain containing protein that can bind with high affinity to negatively charged materials.103 The enzyme is predominantly found in the cells of the myeloid lineage such as neutrophils, and macrophages although to a lesser extent. When MPO is required by these cells, it is released from cytoplasmic granules. The enzyme interacts with H2O2, to form an oxidative complex that can oxidize various chemicals that go on to produce further reactive species. For example chloride (Cl-) can be oxidized first to hypochlorous acid (HClO/ClO-), which can then go on to form chlorine and chloramines.93, 104 Besides having a role in various biological processes including the elimination of pathogens, these powerful MPO products could oxidise CNMs and thereby affect their biodegradation. Several studies affirming this hypothesis have been completed.

For example, one study has shown that fullerenes can undergo biodegradation mediated by - 83 MPO in the presence of 50 μM H2O2 and Cl , over 3 days. The biodegradation was thought to be due to the oxidative action of HClO / ClO-. As part of this oxidation, the hybridisation state of the fullerene carbon atoms would be expected to favourably change from sp2 to sp3 which would help to relieve strain in the structure.23 This would result in the progressive opening of the highly curved truncated icosahedron cage, to form intermediate polyaromatic oxidised species and potentially CO2, although this was not directly evaluated in this study. Oxidised CNHs also have been shown to undergo MPO mediated biodegradation in the - presence of H2O2 and Cl . In the case of oxidised CNHs, we demonstrated that dahlia-like CNH structures were biodegraded over 5 days as evidenced by UV-vis and TEM.105 In terms of kinetics, it appears that the initial dahlia-like super structures of the oxidised CNHs slow down their biodegradation, possibly via retardation of MPO binding. However, when the super structures start to break apart, CNH biodegradation kinetics is faster and close to the SWNT one.

- MPO (with H2O2 and Cl ) has also been shown to efficiently biodegrade carboxylated 84 SWNTs to short-chain carboxylated alkanes, alkenes and CO2. The biodegradation could be inhibited by coating the carboxylated SWNTs with uncharged surfactants, while negatively charged surfactants promoted biodegradation. This was in spite of the fact that both surfactants helped to better disperse the CNMs. Computer simulations revealed that indeed negatively charged surfaces (such as carboxylate groups or negatively charged surfactants) could favour the interaction of the positively charged residues of the enzyme, enabling the oxidative biodegradation of the CNTs. Pristine nanotubes alone with MPO (with - 84 H2O2 and Cl ) were shown to undergo some degree of biodegradation. This may be due to the initial oxidation of the pristine SWNT surface by HClO / ClO- that could facilitate favourable interactions of the enzyme and promote biodegradation. It is evident that by changing the surface characteristics, particularly the charge, of the subject nanostructures through either covalent or non-covalent strategies, one can alter their predisposition towards MPO mediated biodegradation.

This concept was expanded by Bhattacharya et al. who demonstrated that the functionalisation of carboxylated SWNTs with PEG chains (i.e. PEGylation) of increasing molecular weight can in turn increasingly delay their biodegradability.106 PEG groups are hydrophilic uncharged and coiled polymeric chains known to reduce the interaction of proteins and other molecules with nanostructures.107 Therefore, PEGylation of the SWNTs

52 could reduce their ability to interact with enzymes including MPO, hence hampering their ability to biodegrade the CNM.106 By critically analysing the overall findings of these papers, careful designing of the CNMs surface via the grafting of PEG chains for example can help to design the biodegradation behaviours of CNMs by helping to regulate their interactions with enzymes such as MPO.

In addition to conventional surface grafting molecules such as PEG chains, less defined though more biologically relevant molecules such as proteins, have been importantly demonstrated to affect the enzymatic mediated biodegradability of CNMs.86,108 The presence of serum albumin was shown to reduce the kinetics of MPO mediated biodegradation of the carboxylated SWNTs compared to uncoated carboxylated SWNT.86 This observation could be attributed to the competitive binding of serum albumin and MPO to the CNM’s surface which ultimately could reduce the likelihood and frequency at which MPO could interact with the surface of carboxylated SWNTs and therefore reduce the biodegradation kinetics. Moreover, using coatings made of immunoglobin G (IgG), it was shown that immune relevant proteins could impair MPO mediated biodegradability of carboxylated SWNTs109, indicating that this effect is not just limited to albumin but applies to other proteins too. Both of these observations highlight that the presence of protein must be accounted for where MPO mediated biodegradation of CNMs is desired. This is particularly relevant to the biological milieu as well as environmental contexts in which proteins and other macromolecules can be abundant.110, 111 Indeed such findings have also been presented and discussed in the context CNM of biodegradation mediated by isolated chemicals such as NaClO86, which indicates a level of commonality in the factors that can affect the kinetics of non-living systems mediated CNM biodegradation.

As for MPO mediated GO biodegradation, fewer studies have been completed. However recently it has been shown that increasing the dispersibility of the GO sheets, via surface oxidation, can increase the kinetics of biodegradation as catalysed by MPO in the presence - 59 of H2O2 and Cl (Figure 2.12). This is most likely because of the increased surface area available for the MPO to bind to and catalyse GO biodegradation. Accordingly, where rapid biodegradation is desired, GO sheets should exist as highly oxidised sheets of single to few layers that are well dispersed. If prolonged resistance is necessary, the oxidation extent should be reduced or the sheets made thicker.

Figure 2.12 TEM images of hMPO-mediated degradation which evidences that MPO mediated degradation of GO sheets can be regulated by the dispersibility and aggregation state of the material.

The state of the flakes has been captured at 3 time points: control (t = 0), t = 15 h, and t = 24 h post incubation with MPO. Images (A–C) correspond to single-layer well dispersed GO, while (D–F) correspond to intermediately dispersed GO, and G–I) correspond poorly dispersed and highly aggregated GO. Scale bars represent 500 nm. Black arrows (B and E) point to holes within the respective GO flakes after 15 h; white arrows in (D) and (F) mark rod-shaped particles in GO flakes before and after the degradation, and dotted black arrow in (I) indicates degraded GO 3 flakes. Adapted from Dispersibility-Dependent Biodegradation of Graphene Oxide by Myeloperoxidase, 11, Rajendra Kurapati, Julie Russier, Marco A. Squillaci, Emanuele Treossi,nCécilia Ménard-Moyon, Antonio Esaú Del Rio-Castillo, Ester Vazquez,Paolo Samorì,Vincenzo Palermo, Alberto Bianco, Copyright (2015), with permission from John Wiley & Sons, Inc.

2.6.4. Other enzymes.

Other than MPO, other mammalian peroxidase enzymes have been explored including eosinophil peroxidase (EPO). EPO has 68 % sequence identity to MPO and is also a Fe-

54 containing haloperoxidase.112 The enzyme was shown to biodegrade carboxylated SWNTs - 113 over 5 days in the presence of H2O2 and Br , which it uses to produce hypobromous acid. The Cl- was not used here as EPO, unlike MPO, preferentially uses Br-. Using computer modelling, two potential binding sites were identified by which EPO could interact with oxidised SWNTs and catalyse their biodegradation, and like in the case of MPO, they involve the negatively charged carboxylate groups of carboxylated SWNTs.112 This once again highlights the importance of surface properties, particularly charge, of CNMs with regards to biodegradation and illustrates that surface properties may be key to defining CNM biodegradation kinetics at least at the isolated chemical and enzymatic level. EPO mediated biodegradation was found to be more efficient at acidic pH values113, suggesting that the particular pH environment in which degradation occurs can impact the enzyme mediated biodegradation kinetics of oxidised SWNTs and potentially other CNMs. This is crucial as pH varies temporally and spatially in physiology114, 115 as well as in other scenarios where biodegradation is relevant such as the environment.116 It is well known that all enzymes are particularly sensitive to pH, and optimal pH values exist for each. The pH of the target environment should therefore be considered when designing biomedical applications in healthy vs. inflamed or diseased states.

In addition to members of the peroxidase family, other enzymes such as inducible nitrogen oxide synthase (iNOS) and nicotinamide adenine dinucleotide phosphate (NADPH) oxidase have been implicated in the biodegradation of CNMs, particularly CNTs.117, 118 Both these enzymes are not required to bind to the CNMs in order to catalyse biodegradation. Rather they produce radicals such as reactive oxygen or reactive nitrogen species (ROS and RNS, respectively), which can then diffuse over short distance towards the adjacent targets and promote their oxidative biodegradation. The factors that could influence the biodegradation kinetics of CNMs under these circumstances would more closely relate to the factors that influence the chemical degradation of these materials discussed in Section 2.5. These free radicals are small species and therefore will be less restricted by steric hindrances imposed by nanomaterials, such as due to the restrictive inner core diameters of CNTs, and the need for specific binding orientations. This was demonstrated by in situ TEM, highlighting the •OH generation which could biodegrade MWNTs from both internal and external surfaces.117

In addition to biomedically relevant enzymes, enzyme based investigations have been conducted to specifically investigate the biodegradability of CNMs in the environment. Although not the focus of this review, these studies can help to glean further understanding

55 concerning the mechanistics of CNM biodegradation as a whole. The enzymes explored in these studies include non-mammalian enzymes such as lignin peroxidase (LiP) and manganese peroxidase (MnP), both of which also happen to be members of the peroxidase family.119 These enzymes are produced by ubiquitous fungal, mushroom and bacterial species, and are naturally able to biodegrade polyaromatic hydrocarbon and other environment pollutants.120

For example, Chandrasekaran et al. demonstrated that lignin peroxidase (LiP), isolated and partially purified from the fruiting bodies of the mushroom Sparassis latifolia, could efficiently biodegrade carboxylated SWNTs to low molecular weight molecules and eventually CO2, in 121 the presence of H2O2. However, the enzyme was unable to biodegrade pristine SWNTs. In contrast, the same enzyme was unable to catalyse the biodegradation of other carboxylated SWNTs when extracted from the white-rot fungus.122 This difference may be on account of species specific differences or material characteristics, emphasising the need for careful understanding of the biodegradation system employed and the physicochemical differences between materials baring the same name. LiP has also been shown to catalyse the facile degradation of oxidised graphene nanoribbons within 92 h in the presence of H2O2 and the cofactor veratryl alcohol (VA).123 Biodegradation did occur in the absence of VA but at a much slower rate. The same study showed that under the equivalent conditions, the biodegradation of reduced graphene nanoribbons was heavily attenuated. As for the HRP and mammalian peroxidase enzymes, surface charge and chemistry appear to be crucial factors that generally influence the biodegradation of CNMs by LiP. In addition, the capability of an enzyme to biodegrade a material should not be accepted blindly. The source of the enzyme should be considered carefully as species differences may play a part in determining biodegradability.

Manganese peroxidase (MnP), derived from the white-rot fungus, Phanerochaete chrysosporium was capable of bio-transforming pristine SWNTs.122 Contrary to other investigations, it was unable to biodegrade the carboxylated SWNTs, in contrast to LiP, also tested in the same study. The authors hypothesised that the carboxylate groups of the SWNTs could inhibit the action of the MnP enzyme by complexing with the enzyme Mn2+ within the enzymatic binding site. This complexation is thought to be unfavourable and could lead to the inhibition of one or more critical steps in the MnP catalytic cycle. Although this is the first time such an unusual result has been reported, it serves to demonstrate that the features of CNMs that promote enzyme mediated biodegradation may not be universal, especially with regards to environmental breakdown. Nonetheless, it is unlikely that just one enzyme will be present in the environment and the materials will interact with an array of different enzymes. But when considering breakdown of the materials for waste disposal

56 purposes, the enzyme used could be tailored to the characteristics of the material to allow for facile biodegradation.

2.6.5. Overview of enzyme mediated CNM biodegradation and the lessons learnt.

The enzymatic biodegradation of CNMs can be enhanced by using thin (few layers or single- walled structures) and well dispersed materials that are negatively charged. Materials should not be covalently or non-covalently functionalised with structures such as proteins or PEG chains, which can impede enzyme interactions. The exception to this is where the particular structure is an enzyme substrate that is specifically designed to promote biodegradation. Moreover, the environment in which biodegradation takes place should be optimal to suit both the material’s dispersibility as well as the enzyme activity. If slow biodegradation kinetics is desired, the contrary of these statements should be employed (Figure 2.13). Though we give some general guidelines, the specific system should be first characterised, since there will be some exceptions to the rule, for example in the case of manganese peroxidase122. This added level of complexity does not need to be considered in chemically mediated degradation using harsh oxidants because they lack the requirement for specific protein tertiary and quaternary conformations as well as specific binding orientations with the exception of certain enzymes such as NADPH oxidase and iNOS which directly produce reactive species that can then diffuse over short distances and oxidise target nanostructures. We have provided an exhaustive table which lists all studies which have focused on the enzymatic biodegradation of CNM (Table 2.2).

Enzyme specific substrates Non specific coatings Optimal conditions for enzyme (temperature, pH, cofactors Enzyme mediated etc.) CNM biodegradation Aggregation Kinetics

Few layers (walls) PEGylation

Highly oxidised surface (with some exceptions)

Figure 2.13 A summary of the factors that can influence the kinetics of enzymatically mediated CNM biodegradation.

Where green arrows indicate acceleratory influences on the kinetics (indicated by the + symbols) while red arrows indicate retarding factors on the kinetics (indicated by the – symbols).

Table 2.2 A list of the literature studies that have specifically focused on the enzyme mediated degradation of CNMs.

Type Modification Model Summary of Finding Reference

95Allen et al., 2008) SWNT Ox HRP  Evidence provided of the degradation of CNTs by HRP/H2O2.

H2O2 + Fe

SWNT Pristine HRP  HRP is unable to degrade pristine SWNT unlike carboxylated SWNT which degraded over time. 96

 CO2 was final product. Ox H2O2 + Fe  Oxidised poly aromatic hydrocarbons are by-products of degradation. Hemin + Fe

84 SWNT Ox MPO +H2O2 and Cl-  Hypochlorite and reactive radical intermediates of the human neutrophil enzyme myeloperoxidase induces the degradation of single-walled CNTs in neutrophils and to a lesser degree in macrophages. IgG coated  IgG-functionalised nanotubes induce the release of MPO and the generation of reactive oxygen species in human peripheral blood.

MWNT Ox (5 h) HRP  Length and diameter reduced. 99  Layers removed in a sequential manner. Ox (8 h)  Nitrogen doped MWNT degraded to a greater extent than COOH functionalised MWNTs. Nitrogen doped

124 SWNT Ox MPO  Degradation was possible due to the presence of HOCl produced by MPO and HOBr produced by LPO LPO.

SWNT Ox PSF  SWNT were degraded to a greater extent than MWNT. 88  The tubular structure of MWNT could be identified after 60 days under both PSF and HRP induced MWNT HRP degradation.  The presence of defects encourages degradation.  HRP is more efficient than PSF.

Graphene Ox HRP  HRP catalysed the degradation of graphene oxide and hole formation was indicated via electron 101 microscopy.

 Computational analysis revealed binding sites on HRP to interact with GO.  The heme active site of HRP was in closer proximity, and the enzyme itself was more dynamic when bound to GO compared to rGO and so GO is more susceptible to degradation

 CO2 was seen to be the final end product.

125 SWNT Ox HRP + H2O2  Maximum DNA damage was induced by degradation products obtained on day 7 of HRP mediated degradation of SWNT.

SWNT Ox EPO  Human EPO (in vitro) and murine EPO from ex vivo activated eosinophils catalyses the oxidative 113 degradation of SWCNTs.

SWNT Ox LiP  Over 20 days, both thermally treated and raw carboxylated SWNT were efficiently degraded by lignin 121 peroxidase, but pristine SWNT were not. Pristine  SWNTs were mineralised to CO2

SWNT Pristine LiP,  LiP and Lacc were unable significantly degrade carboxylated or pristine SWNT over a period of 16 122 days. MnP  Manganese peroxidase was able to induce defects to pristine but not carboxylated SWNTs over a Laccase period of 16 days.

SWNT Ox MPO  MPO was able to degrade functionalised SWCNTs over 7 days. Oxidised non-PEGylated SWNT 106 underwent the fastest degradation, as PEG molecular weight increased, the rate of degradation PEG decreased.  Primary human neutrophils were able to degrade oxidised and PEGylated SWNTs over 8 h both intra and extra cellularly. The degradation efficiency was independent of functionalisation strategy and PEG molecular weight.

MWNT Ox HRP  Coumarin and catechol moieties have enhanced the degradation of MWCNTs by HRP compared to 100 oxidised nanotubes. 3-azido-7-hydroxycoumarin XO  Purine-functionalised MWNT underwent degradation by XO although at a reduced rate compared to 4-methyl-7-(azidopropyl) oxidised nanotubes. coumarin

9-(3-azidopropyl)-purine

3,4-dihydroxybenzoic acid

Fullerene Pristine NaClO  MPO was able to degrade C60 fullerene as evidenced with UV-vis spectroscopy and FT-IR. 83

MPO

HRP = Horse radish peroxidase, MPO = Myeloperoxidase, EPO = Eosinophil peroxidase, LPO = Lactoperoxidase, LiP= Lignin peroxidase, MnP = Mangansese 60 peroxidase, XO = Xanthine oxidase Ox = oxidised or carboxylated.

2.7. Biodegradation mediated by cells.

When considering the interactions of CNMs with living systems in general, or the engineering of materials for biomedical applications, it is critical that the biodegradability or biopersistence and the kinetics of these processes are understood in the context of true biological environments and their complexity. Within cells, enzymes are heavily relied upon to catalyse various essential processes from regulation of the genetic materials to the recycling of cellular waste. Research discussed in Section 2.6 provides compelling evidence to demonstrate the capacity of some of enzymes to biodegrade CNMs. As a result, over the last few years there has been a growing research effort aimed to demonstrate that CNM biodegradation can also occur when CNMs interact with living cells.

2.7.1. Neutrophils.

The first cells to be investigated in this regard were neutrophils. Neutrophils are mobile short-lived phagocytic cells that function as one of the main effector cell types of the innate immune system.126, 127 The purpose of this immune cell type includes, but is not limited to, the rapid localisation, phagocytosis and subsequent destruction of invading pathogens and exogenous materials.126 Neutrophils are typically the first cells to arrive at sites of infection. Besides this function, they also work to shape the overall immune response and tissue repair processes.128 These cells contain a large amount of MPO stored within cytoplasmic granules. When activated, they can release MPO both intracellularly and extracellularly via degranulation, hence inducing death of bacteria and other pathogens.104, 129 The cells can also be stimulated to produce neutrophil extracellular traps (NETs) which are rich in MPO during a process known as NETosis.130 During NETosis, the neutrophil’s nucleus decondenses and the nuclear envelope ruptures. This results in the mixing of various cellular components such as chromatin and granular components. The rupturing of the cell membrane then occurs with the consequent extrusion of the cellular components, which include significant quantities of oxidative enzymes such as MPO.130 Therefore neutrophils are promising candidates regarding the biodegradation of CNMs.

Kagan et al. were the first to report that carboxylated SWCNTs can undergo MPO-mediated biodegradation within isolated human neutrophils under in vitro conditions.84 In order to induce the MPO expression, the neutrophils were activated with N-formyl-methionylleucyl-

61 phenylalanine, whilst cytochalasin B was employed as a degranulation promoting agent to stimulate the release of MPO from neutrophil cytoplasmic granules. When the cells were incubated with a suspension of carboxylated SWNTs over 12 h, only 30 % of the material underwent biodegradation. It was found that the majority of the carboxylated SWNTs had not been internalised by the cells, and so biodegradation even with extracellular release of MPO was inefficient. Therefore to increase their internalisation in the cytoplasm where MPO concentration is higher, the carboxylated SWNTs were first functionalised with IgG to promote receptor mediated endocytosis. When the biodegradation of IgG-carboxylated SWNTs was investigated under the same conditions, the materials underwent 100 % biodegradation.84 This study demonstrates that internalisation is of key importance with regards to CNM biodegradation by neutrophils under these circumstances. The positive influence of IgG on the uptake and thus the biodegradative capacity of neutrophils has recently been confirmed by another group of researchers.109 The same team also demonstrated a similar uptake and biodegradation enhancement when bovine serum albumin (BSA) was used to coat the SWNTs.108 Moreover, when the BSA-coated SWNTs were injected intraperitoneally, activation occurred in vivo; and in addition to neutrophils, macrophages were also activated. This study emphasises the important role that macromolecules attached to CNMs can have on the biodegradative capacity of cells implicated in their destruction.108 This is in strong contrast to acellular in vitro enzyme and chemical mediated CNM biodegradation, where coatings such as IgG and albumin108, 109 and other molecular coatings such as PEG106 actually hinder biodegradation by MPO enzymes alone. Such observations can be made because unlike acellular in vitro systems, cells are able to respond and change their behaviour in response to certain molecular cues, such as IgG, by increasing the availability of MPO as well as other enzymes. This has been portrayed in Figure 2.14.

Figure 2.14 Proposed mechanisms for IgG coating-dependent degradation of SWNTs in neutrophils.

Reprinted from Effects of serum albumin on the degradation and cytotoxicity of single-walled carbon nanotubes, 222, Yun Ding,Rong Tian,Zhen Yang,Jianfa Chen,Naihao Lu, Pages No 1 - 6., Copyright (2016), with permission from Elsevier.

Together these studies indicate that the protein coronation of CNMs could be used to tailor their biodegradation kinetics for a number of applications. To further understand this phenomenon, studies should be conducted to determine if the same enhancement of biodegradation can occur following the incubation of CNMs in more biologically relevant media such as blood plasma, where serum albumin is one of the most abundant proteins.131 This is particularly important with regards to the design of intravenously administered drug delivery vectors and diagnostic agents, which will encounter and bind plasma proteins that could subsequently cause activation of neutrophils or macrophages. Such factors could affect the biodegradation kinetics of the materials, and would not necessarily be predicted via acellular in vitro enzyme experiments.

Recently, it has been shown that besides biodegrading CNMs intracellularly, neutrophils can promote biodegradation extracellularly132 via the stimulated production of NETs that, as mentioned, are filled with biodegradative enzymes such as MPO. When promoted to do so

63 by specific stimuli, these NETs were able to trap and biodegrade carboxylated SWNTs as shown in Figure 2.15. This further demonstrates that the diverse range of responses and behaviours that cells display can influence their biodegradative capacity.

i ii iii

iv v vi

Figure 2.15 NET MPO mediated biodegradation of SWNTs.

(A) TEM images of SWNTs after 24 h incubation under the following conditions: (i) SWNTs in H2O; (ii)

SWNTs + NETs; (iii) SWNTs in H2O2 alone; (iv) SWNTs in H2O2 + NaBr; (v and vi) SWNTs + NETs supplemented with H2O2 and NaBr (B) UV-vis/NIR spectroscopy assessment of biodegradation. The intensity of the S2 absorbance spectra (981 nm) of SWNTs treated as indicated at 0 h (black bars) and 24 h (white bars). The reduced intensity of the S2 absorbance spectra (white bars) represents degradation of the SWNTs in the suspension as seen after 24 h treatment. Adapted from C. Farrera, K. Bhattacharya, B. Lazzaretto, F. T. Andon, K. Hultenby, G. P. Kotchey, A. Star and B. Fadeel, Nanoscale, 2014, 6, 6974-6983 with permission of The Royal Society of Chemistry.

Therefore with regards to neutrophils, in addition to the type of enzymes those cells contain, the particular activation state the cells are in can greatly influence their ability and the kinetics with which they can biodegrade CNMs. This activation state can in return be modified by the specific engineering of the biological corona that surrounds the CNMs. The binding of protein can occur in a controlled fashion at the CNM design stage by chemical functionalisation (i.e. attended corona), or could occur spontaneously in normal cell culture environment or in experimental animals albeit in a less well understood or controlled manner (i.e. unattended corona).111, 133 In both cases, these strategies can be engineered to favour or inhibit biodegradation of CNMs.

Despite the capacity of neutrophils to effect the biodegradation of the 1D functionalised CNTs, there are currently no studies exploring their potential to biodegrade fullerenes, carbon nanohorns or graphene. Therefore further studies are warranted in this area.

2.7.2. Macrophages.

Besides neutrophils, other professional phagocytic cells such as mobile or tissue fixed macrophages134 may have a role in the biodegradation of CNMs. Macrophages are derived from bone marrow stem cells and migrate into tissues as circulating monocytes before differentiating into mature macrophages.135 There is also evidence to suggest that tissues such as the heart and the liver may contain tissue-specific macrophages that are initially derived from the embryonic yolk sac during foetal development and persist into adult hood.136 Macrophages are long lived cells that are involved in the phagocytosis and the elimination of dead cells as well as cellular debris. For instance, macrophages contribute to specific processes such as the destruction of damaged erythrocytes, the recycling of Fe, tissue remodelling, as well as antigen presentation.137 Moreover, they are key players in the innate immune system, where they can phagocytose and destroy invading microorganisms such as bacteria, as well as exogenous particulate matter.134 Unlike neutrophils, macrophages express lower quantities of oxidising enzymes such as MPO, NADPH oxidase and iNOS.

Kagan and co-workers were the first to demonstrate that macrophages can biodegrade IgG- coated carboxylated SWNTs, with reduced kinetics in comparison to neutrophils84. In another study, the team also highlighted the role that iNOS and membrane-bound enzyme NADPH oxidase play in macrophage mediated biodegradation of carboxylated SWNTs.138 It was later shown that carboxylated MWNTs can also be internalised in macrophages and

65 begin to undergo a slow biodegradative process.117 Biodegradation in this case was attributed to a combination of intracellular enzymatic systems, including MPO, iNOS NOX and NADPH oxidase.

The mechanism of macrophage mediated biodegradation was further explored by Ding et al. 118, who showed that carboxylated SWNTs underwent significant though not complete intracellular biodegradation in zymosan-activated macrophages over 48 h. Biodegradation did not occur when the macrophages had not been priorly activated. Zymosan activation of the macrophages resulted in an upregulation of the production of MPO, NADPH oxidase and iNOS; all of which have shown capability to biodegrade CNMs. As with neutrophils84, 109, functionalising the carboxylated SWNTs with IgG resulted in activation of the macrophages as well as a better phagocytosis, resulting in a more rapid biodegradation of IgG-coated vs. non-coated carboxylated SWNTs.118

Therefore, the activation state of the cells and the presence of molecular cues on the CNMs appear, as for neutrophils, to be essential factors that can modify the responses of macrophages thereby changing the kinetics of CNM biodegradation. When designing CNMs for biomedical applications, careful adaptation of the specific characteristics of the CNM structure in relation to the cells and the conditions of the microenvironment will hence need to be considered. Indeed, there is evidence to suggest that changes to the surface chemistry to CNMs can in turn lead to changes in the abundance of various protein components of their biomolecule corona, such as IgG, that develop in biological contexts.139 This in turn can have major influences on the neutrophil responses to the materials and therefore its ability and the kinetics by which it could biodegrade them, as described.

There is also evidence to suggest that the presence of Fe within CNTs, can delay macrophage mediated intracellular CNT biodegradation compared when no Fe is present.140 In this study, the authors suggested that following internalisation of the Fe@CNT, the macrophages would first be required to alter their intracellular free Fe concentration on account of the added Fe content coming from the CNTs. As a consequence of this, there was an initial delay in the CNT biodegradation process in comparison to Fe-free CNTs. Considering that various CNMs including CNTs and graphene can be contaminated with metal impurities including Fe141, when one is designing materials with tailored biodegradation kinetics, the purity of the materials should also be considered.

In addition to their Fe content, the surface properties and geometries of MWNTs can curiously modulate macrophage intracellular processes and thereby influence their ability to mediate MWNT biodegradation. Landry et al. showed that the collective influence of MWNT length and oxygen functionalisation effects macrophage intracellular pH.142 This was such that the exposure of macrophages to short pristine MWNTs and long MWNTs (pristine or oxidatively functionalised) triggered a decrease in cellular pH, which was associated with early MWNTs biodegradative changes, with short pristine MWNTs demonstrating the greatest biodegradative changes. On the other hand, short functionalized CNT did not induce this early intracellular pH decrease and were shown to be immune to early biodegradative changes.142 On account of the early time points investigated (up to 48 hours) and the inactivity of enzymes such as MPO and NADPH oxidase at low pH143, 144, the intracellular pH related biodegradation of MWNT may be linked to cellular functions that specifically rely on a decrease in intracellular pH. Further investigations are required to determine these details. This study reaffirms the importance of considering CNMs physicochemical characteristics when considering their biodegradation kinetics. These characteristics appear to be equally important at the initial stages of biodegradation, as it is over the course of the whole process and therefore should be considered. Moreover, this study emphasises that biodegradation processes may not be simple but could be the result of a cascade of events; each of which could be regulated in different ways.

In addition to the studies carried out for CNTs, macrophages have been demonstrated to biodegrade both carboxylated graphene sheets145 and oxidised CNHs.105 It is likely that these two types of materials could presumably undergo biodegradation with increased kinetics compared to CNTs. This is on account of the combination of higher surface area for enzymes to bind to and generated free radicals to react with. Moreover in oxidised and defected materials, a higher surface area means a higher number of available structural defects at the surface that could serve as biodegradation initiator sites.63

2.7.3. Microglia.

Microglia, the resident macrophages of the central nervous system (CNS) have started to be explored for their propensity to biodegrade CNMs, specifically CNTs. Microglial cells are of mesodermal/mesenchymal origin and persist into adulthood. They are present in all regions of the CNS and are spread throughout the brain parenchyma.136, 146 In their resting state, microglia are characterised by highly motile processes which they use to survey their local

67 environment for signs of pathology and debris.147 Upon detection of such a stimulus, resting microglial cells can undergo a multistage activation that transforms them into the activated microglial cell.148 Unlike neutrophils and macrophages, healthy microglial cells hardly express MPO149, 150, however they do express other enzymes such as iNOS151 and NADPH oxidase.152 The production of these enzymes is upregulated when the cells are activated.

Goode et al. demonstrated that the specific interaction of microglial cells with MWNTs in vitro crucially depends on the oxidation state of the surface of the MWNTs.153 Microglial cells appeared to slowly dismantle large agglomerates of carboxylated MWNTs and subsequently internalise smaller fragments. In contrast, the larger and more tightly bound agglomerates of pristine MWNTs seemed to remain predominantly intact throughout the full experiment, despite their close association with microglial cells. On rare occasions however, sufficiently small agglomerates of pristine MWNTs were subject to microglial uptake. When the interior of the cells were observed via TEM, oxidised MWNTs could be identified within endosomes, multilamella bodies as well as free within the cytoplasm. The oxidised MWNTs were shown to have undergone morphological changes that were not evident for the pristine MWNTs. These changes included delamination and disordering of the graphitic structure of the outer walls. It was not certain however, whether this was due to damage during initial oxidation during oxidative synthesis procedures or due to cell mediated biodegradation of the oxidised MWNTs outer walls. Although this study did not definitively demonstrate the biodegradation of MWNTs within microglial cells, it shows that the surface oxidation state of MWNTs has vital implications concerning the propensity of microglial cells to disentangle and uptake oxidised MWNT agglomerates. These two parameters would be required for efficient microglial mediated biodegradation.

Therefore, when incorporating CNMs such as MWNTs for biomedical applications, the success of which may partially depend on specific biodegradation kinetics of the materials which in turn is regulated by the oxidation state as well as the dispersibility of the MWNTs. Hence these parameters must be carefully tuned. This key consideration was highlighted in two reports: firstly by Mata et al., when investigating the in vivo biodegradability and biocompatibility of CNT based scaffolds154, and secondly by Kurapati et al. when exploring the MPO mediated degradation of GO.59

Beyond oxidation state, differences in the specific types of surface chemistry could also affect CNM biodegradation kinetics mediated by cells. This was tested in microglial cell

68 cultures, where the biodegradation kinetics of carboxylated, aminated, and both carboxylated and aminated MWNTs were assessed over 3 months.155 Biodegradation was identified in all three cases, suggesting that the microglial mediated biodegradability of functionalised MWNTs could be independent of surface chemistry. However, the two types of carboxylated MWNTs seemed to follow a faster rate of biodegradation over the first 2 weeks of the study. Beyond this time point, all three types appeared to have similar biodegradation kinetics (Figure 2.16). This was interesting since it was the first time that aminated MWNTs were explored for their biodegradability in vitro. Therefore, enhancement of the biodegradation kinetics of MWNT mediated by microglia not only can be attained by oxidative functionalisation, but other strategies can also been employed including amination, which is important with regards to some therapeutic applications of MWNTs, for example in the therapeutic nucleic acid delivery strategies.156, 157

Figure 2.16 Raman spectroscopy of microglia cell culture monolayers after exposure to different functionalised MWNTs.

Following the collection of a Raman spectrum, the intensity ratio between the D and G bands was calculated from the normalised intensities of the D and G bands for each spectrum. Intensities were normalised to the intensity of G band. The average ID/IG band intensity ratio (A, B, C) are presented for microglia exposed to (A) ox-MWNT, (B) ox-MWNT-NH3 + and to (C) MWNT-NH3 +. An overall continuous decrease of ID/IG band ratio was observed over time, more significant during the first 14 days for both oxidised materials compared to aminated MWNTs. Between 12 and 22 spectra were collected and analysed for each sample. Adapted from C. Bussy, C. Hadad, M. Prato, A. Bianco and K. Kostarelos, Nanoscale, 2016, 8, 590-601 with permission of The Royal Society of Chemistry.

2.7.4. Overview of cell mediated CNM biodegradation and the lessons learnt.

Overall, the factors that influence the biodegradation kinetics of CNMs are more complex and diverse within cells compared to simple interaction with isolated enzymes or chemicals. This is on account of cells’ ability to respond to various molecular cues by changing their activity in responses to stimuli. Moreover, based on the reports published so far, functionalisation seems to promote biodegradation; however the type of surface functionalisation does not seem to significantly influence kinetics except at the earliest stages of the biodegradation process. This might be due to intracellular substances (protein, lipids, etc.) that may coat the nanomaterial’s surface, modulating the effects of surface charge. Moreover, within a cell, it is likely that a combination of enzymes and other biodegradative conditions will be involved in the material’s biodegradation. We summarise the current understanding of the factors that can influence biodegradation in Figure 2.17 and provide an exhaustive table (Table 2.3) of all the published studies that have specifically explored the cell mediated biodegradation of CNMs. So far, knowledge is limited to only a few studies, therefore it is difficult to draw definitive conclusions with respect to the identities of all cellular components that contribute to CNM biodegradation. More work is hence needed to supplement the current knowledge allowing a full understanding of the cell based causative agents that mediate CNM biodegradation.

Cellular activation

Multiple Protein walls coronation

Good Cell mediated CNM dispersibility biodegradation Aggregation Kinetics

Covalent Functionalisation Iron

Targeting peptides

Figure 2.17. Summary of the factors that can influence the kinetics of cell mediated CNM biodegradation.

Where green arrows indicate acceleratory influences on the kinetics (indicated by the + symbol) while red arrows indicate retarding factors on the kinetics (indicated by the – symbol).

Table 2.3 A list of literature studies that have specifically focused on the cell mediated degradation of CNMs.

Type Modification Model Summary of Findings Reference

MWNT Ox PC-3  Nanotubes were most likely to have been exocytosed. 158

HeLA

SWNT Ox NaClO  The formation of an albumin corona resulted in reduced degradation kinetic of SWNT following 86 exposure to OCl− or human MPO in vitro. Albumin MPO  Albumin−SWCNT interactions enhanced the cellular uptake of SWNTs and stimulated neutrophils to Neutrophil upregulate production of MPO.  Upon neutrophil stimulation, both SWCNTs and albumin-SWCNTs were significantly biodegraded. Although, the degradation of albumin-SWNT was more evident.

SWNT Ox Primary Neutrophils  On incubation of SWNT with neutrophils, neutrophil extracellular traps can be stimulated to form 132 (stimulated by phorbol 12-myristate 13-acetate) which then trapped clusters of aggregated SWNT.  SWNTs underwent significant degradation when incubated with the isolated of neutrophil extracellular traps and NaBr over 24 h.  Addition of Proteinase K eliminated enzymatic activity of the NETs and this also reduced the

degradation of SWCNTs in the presence of H2O2.  Incubation of partially degraded SWNT with freshly isolated human neutrophils resulted in significant cytotoxicity after 6 h.

CNH Ox HRP  On incubation of CNHs with MPO, 60 % of the material was degraded within 24 h. 105  When exposed to cells in vitro, ~ 30 % of the CNHs were degraded within 9 days. RAW 264.7 and THP-1 macrophage cells

117 MWNT NH3 Macrophages derived from  After 168 h of exposing macrophages to MWNT, the cells had internalised and created holes in the THP-1 monocytes MWNT; there was also evidence of walls exfoliation.  The extent of degradation was attenuated by the addition of ROS scavengers.  Products of degradation were less cytotoxic than the non-degraded MWNTs.  Following exposure of rats to MWNT, MWNT were accumulated within lung and lymph bound macrophages. After 7 days, the MWNT showed evidence of morphological damage (hole formation) consistent with the early stages of degradation.  The holes formation and the thinning of the MWNTs was recreated using in situ TEM derived OH•

radicals.  Molecular and chemical biology analysis of exposed macrophages evidenced the upregulation of NADPH oxidase and superoxide dismutase, myeloperoxidase and inducible nitrogen oxide synthase enzymes.

155 MWNT NH2 Primary microglia  Primary microglia internalised and progressively degraded the functionalised MWNT irrespective of their surface chemistry over a period of 3 months.

SWNT COOH MPO  When mixed with IgG, the protein spontaneously binds to SWNTs. 109

IgG coated Neutrophils  The presence of an IgG protein corona reduced the MPO-mediated SWCNTs biodegradation in vitro.

 IgG-absorbed SWCNTs stimulated neutrophils to produce MPO and OCl−.

 The degradability of IgG-absorbed SWCNTs greater than the non-coated carboxylated CNTs in activated neutrophils.

SWNT Pristine RAW 264.7 macrophages  Macrophages can initiate the degradation of ox-SWNTs and OH-SWNTs but not pristine SWNTs. 159

Ox Full enzyme solution  Degradation process was enhanced via PMA-induced respiratory burst in macrophages.

(MPO + SOD + NOX2) OH  N-Acetyl Cysteine can efficiently block the process macrophage induced degradation.

 No degradation was detected for pristine SWNT using the full enzyme solution over 30 h. This was unlike ox-SWCNTs and OH-SWCNTs; both of which exhibited signs of degradation, with OH-SWNT being degraded to a greater extent than ox-SWNT.

 N-Acetyl Cysteine can efficiently block the process enzyme solution induced degradation.

MWNT S-MWNT (short) RAW 264.7 macrophages  All CNT could be internalised by macrophages, mainly inside vacuoles although to a lesser extent for 142 SF-MWNT. SF-MWNT (short and functionalized),  All MWNTs apart from SF-MWNT induced acidification of the intracellular lysosomal compartment.

L-MWNT (long)  Short pristine CNT were more prone to biodegradation than long CNT (pristine or functionalised), while short functionalised CNT were protected. LF-MWNT (long and functionalized),  Incubation of cells with concanamycin completely prevents CNT from being modified, indicating the

Functionalized CNT (SF- and importance of intracellular pH.

LF-MWNT)

Functionalisation : Ox

SWNT Ox MPO  Binding of serum albumin to SWNTs could reduce the kinetics of MPO mediated degradation due to 108 competitive adsorption. BSA coated Macrophages  Binding of serum albumin to SWNTs enhanced the kinetics of macrophage mediated degradation; this degradation was enhanced during macrophage activation.

 BSA coated SWNTs induced less cytotoxicity than non-coated SWNTs.

 Biodegraded SWNTs were less cytotoxic than non-degraded SWNTs.

140 MWNT NH2 THP-1-derived  Intracellularly-induced structural damages appear more 2.7 times more rapidly for MWNT compared to macrophages fe@ MWNTs, however after 168 h these differences became insignificant. Fe @ MWNT  N-Acetyl Cysteine reduced the degradation kinetics of MWNT to a greater extent than Fe@MWNT.

 CNT exposure activates an oxidative stress-dependent production of iron via Nrf2 nuclear translocation, Ferritin H and Heme oxygenase 1 translation.

 CNT exposure promoted Bach1 to translocate to the nucleus of cells exposed to Fe @ MWNT in order to recycle embedded fe, which ultimately delayed degradation compared to MWNT.

MPO = Myeloperoxidase, SOD = Superoxide dismutase, BSA = Bovine serum albumin, NOX2 = Subunit of nicotinamide adenine dinucleotide phosphate- oxidase, Ox = oxidised or carboxylated.

2.8. Biodegradation of CNMs in vivo.

The next level of complexity after cells that has been explored to investigate the biodegradability of CNMs is animal models. This level of research is among the aims of preclinical studies where researchers seek to obtain proof of concepts and to characterise the pharmacodynamics and pharmacokinetics of intended products in animal models. This often includes the assessment of materials’ in vivo biodegradability and occurs before subsequent phases of clinical trials in human subjects.160 Since the study of the biodegradability of CNMs is still maturing, only a few in vivo studies have been thus far reported.

2.8.1. Biodegradation of CNMs in the lungs.

The first organs to be investigated in the contexts of CNM biodegradability were the lungs161. This is because following the inhalation of airborne nanomaterials; the lungs are the most targeted organ. Moreover the inhalation route is highly relevant to unintended occupational and private exposures. Early suggestions of the in vivo biodegradation of MWNTs (with few oxidative defects) within the lungs came in 2008, following the intra-tracheal administration of the material to rats.161 After 15 days, broncho-alveolar lavage was performed and the recovered MWNTs were characterised. Microscopic analysis showed that the MWNTs have undergone a reduction in length, while spectroscopic analysis suggested that the tubes may have sustained some oxidative changes. However, because the biodegradability of the MWNTs was not the main end point of this study, no biodegradation mechanism was explored.

In contrast, a hypothesis driven investigation of the biodegradation of carboxylated SWNTs in lungs was conducted by Shvedova et al.162 In this study, the oxidised SWNTs were administered via oropharyngeal aspiration to mice which did not express the MPO enzyme (i.e. MPO deficient mice) or to normal MPO competent mice. Strikingly, the administered carboxylated SWNTs had been cleared to a large extent from the lungs of the MPO competent mice, but were still present in the MPO deficient mice after 28 days, suggesting that MPO is important for clearance (i.e. biodegradation). Further investigations of the state of the carboxylated SWNTs using spectroscopy and microscopy, evidenced the

76 biodegradation and morphological changes of the residual SWNTs in MPO competent mice (Figure 2.18). The persistence of the materials within the lungs of the MPO deficient mice was associated with a robust and persistent inflammatory response that did not occur in the exposed MPO competent mice. The result highlights the role of biodegradation in the elimination of pro-inflammatory exogenous materials. This study was the first to confirm the vital role of MPO and neutrophils in the oxidative biodegradation of CNMs within the lungs. In a follow-up study, the same team showed that 28 days after pulmonary exposure of mice to carboxylated SWNTs, most of the materials had been cleared from both the lung parenchyma and the alveolar macrophages that had previously entrapped most of the lung load of the CNTs after the initial neutrophil intervention.138 In this case, biodegradation was in part attributed to the activity of macrophage NADPH oxidase via superoxide/ NO* peroxynitrite driven oxidative pathways. Taken together, these investigations reveal that a multiplicity of cells and enzymatic mechanisms are responsible for the rapid clearance of carboxylated SWNTs from the lungs.

The fate of pristine CNMs has also been explored in the lungs. Pristine MWNTs were administered in the lungs via their intra-tracheal administration to rats.163 The authors found that after 1 day approximately 30 % of the lung MWNT burden had been cleared via mucociliary clearance. After this however, the quantity of MWNTs present in the lungs did not change and could be detected within the alveolar macrophages for up to 364 days. In another study, when rats were exposed to aminated MWNTs via the pulmonary route, morphological signs of damage to the walls of the aminated MWNTs could be observed after just 7 days.117 Although investigations at longer time points were not completed for the latter study, these observations and the comparison of the two studies support the idea that surface functionalisation is a prerequisite for biodegradation, as suggested before by studies using in vitro models. 84, 88

A B

Figure 2.18. Raman spectroscopic evaluation of ‘‘oxidative’’ defects in SWNT present in the lungs of wild-type (w/t) and MPO knocked-out (k/o) mice at days 1 and 28 post exposure using single point Raman spectroscopy or Raman mapping of different areas within the tissue samples.

(A) D band/ G- band ratios for single point Raman spectra obtained from samples at days 1 and 28 post exposure. * p < 0.05, vs w/t 1 day post exposure, # p < 0.05, vs w/t 28 days post exposure. Inset – typical Raman spectra (excitation at 633 nm) of solubilised lungs of w/t mice at days 1 and 28 post exposure. (B-D) Raman mapping of SWNTs in different areas of the lung sections. Examples of bright-field images are shown with a red box indicating the area where 32632 Raman spectra were acquired. Note that the sizes of the acquired areas were different at day 1 and day 28 (20 μm x 20 μm and 10 μm x10 μm respectively), as significantly smaller SWNT aggregates were detected at day 28 in w/t mice and the scanned areas were adjusted accordingly. (C) Raman maps (excitation at 473 nm excitation) with examples of spectra corresponding to each of the clusters. (D) D-band/ G-band ratios for Raman spectral maps obtained from the lung of w/t and MPO k/o mice at days 1 and 28 post exposure. Reprinted from A. A. Shvedova, A. A. Kapralov, W. H. Feng, E. R. Kisin, A. R. Murray, R. R. Mercer, C. M. St Croix, M. A. Lang, S. C. Watkins, N. V. Konduru, B. L. Allen, J. Conroy, G. P. Kotchey, B. M. Mohamed, A. D. Meade, Y. Volkov, A. Star, B. Fadeel and V. E. Kagan, PloS One, 2012, 7, e30923 with permission.

2.8.2. Biodegradation of CNMs in other tissues.

In vivo biodegradation of CNMs has also been observed in other tissues. In a study relevant to nanomedicine, we reported the early signs of in vivo biodegradation of aminated MWNTs in the brain cortex over a period of two weeks following administration via stereotactic injection.164 Microscopic analysis demonstrated that functionalised MWNTs were endocytosed by microglia and that their multi-tubular structure underwent a biodegradative change which appeared to proceed via a wall exfoliation mechanism. Moreover, spectroscopic analysis was used to reveal the structural evolution of the functionalised MWNTs over time.

Following subcutaneous administration in mice, Sato et al. demonstrated the intracellular biodegradation of oxidised MWNTs over 2 years.165 Specifically, only the small agglomerates of MWNTs that had been internalised by macrophages underwent gradual biodegradation in lysosomes. The majority of the materials however, remained outside of the cells (i.e. in the interstitial tissue) and remained unchanged. This message is reminiscent of the in vitro work conducted by Goode and co-workers, where only materials that underwent internalisation by microglia showed evidence of structural damage.153

As for graphene-based materials, Girish et al. explored the biodegradability of carboxylated graphene sheets following a single intravenous administration to mice.145 On examining the tissues, it appeared that the materials had accumulated as agglomerates within the lungs, liver, spleen and kidneys. The structural integrity of these entrapped agglomerates was then studied over 3 months directly within the tissue of accumulation. During this time, the agglomerates underwent a gradual amorphisation of the carbon structure (i.e. loss of crystallinity on the edges of agglomerates), yet remained physically intact. The spleen bound agglomerates underwent the most rapid change. The materials however were still present after 3 months and had caused considerable toxicity.

The same group recently presented a comprehensive analysis, in which they compared the biodegradability of three materials, namely carboxylated graphene, PEGylated graphene and pristine graphene.166 They demonstrated that irrespective of surface chemistry, graphene- based materials accumulated within the first 24 h primarily in lungs followed by the spleen, liver and kidneys. Significant accumulation of materials in the lungs is expected where the

79 size of the flakes or agglomerates are larger than the diameter of the lungs capillary (5 - 7 μm).47, 55 The lung capillary bed is the first set of capillaries encountered by a nanomaterial following intravenous administration. Spectroscopic analyses revealed that the pristine material maintained its Raman spectral integrity over the 3 months, whereas carboxylated and PEGylated graphene samples showed significant changes, indicative of biodegradation. Unfortunately, the chronic changes and fate of the materials were not traced further and the identities of the cells in which the materials were up taken were not determined. Nevertheless, this study demonstrated that biodegradation of functionalised graphene, as opposed to pristine graphene, may be possible.

2.8.3. Overview of in vivo mediated CNM biodegradation and the lessons learnt.

No further in vivo studies have been conducted to examine the biodegradability of fullerenes, nanohorns or even graphene oxide, which is the derivative of graphene that has attracted considerable attention regarding its biomedical applicability. Overall, it seems that functionalised CNMs are more biodegradable in vivo than pristine materials. However, the types of cells involved and the factors that control the kinetics of biodegradation still need to be more systematically investigated. For instance, more studies directly comparing the biodegradability of CNMs baring different surface chemistries vs. a pristine surface are needed. Based on the investigations using cell models, it is likely that in vivo biodegradation kinetics will be better enhanced by defect inducing surface functionalisation strategies such as carboxylation. In the meantime, we have listed the factors that we have identified so far that positively or negatively impact biodegradation kinetics in vivo (Figure 2.19) and provided an exhaustive table (Table 2.4) of all the in vivo based studies conducted thus far.

Covalent surface Adsorbed functionalisation macromolecules

sp2 Carbon Nanomaterials biodegradation Kinetics in vivo

High PEGylation dispersibility

Figure 2.19 A summary of the factors that can influence the kinetics of CNM biodegradation in vivo.

Where green arrows indicate acceleratory influences on the kinetics (indicated by the + symbol while red arrows indicate retarding factors on the kinetics (indicated by the – symbol).

Table 2.4 A list of literature studies that have specifically focused on the in vivo degradability of CNMs.

Type Modification Model Summary of results Reference

MWNT Pristine Rat lung  Modification of structure (Presence of alcohol, carbonyl and nitrogen function were observed on the 167 MWCNT).  Reduction in length.

SWNT Ox MPO / MPO knockout  Inflammatory response is significantly higher in MPO knockout mice than in wild type mice. 162  Degradation of CNT was demonstrated in wild type mice but was considerably reduced in MPO B6.129X1-MPO (MPO k/o) knockout mice. and wild-type C57Bl/6 mice (w/t)

164 MWNT NH2 Striatum of 6-8 week  MWNT were up taken by microglia cells. C57BL/6 female mice  Within 14 days, intracellular MWNTs began to show evidence of degradation via a wall exfoliation mechanism.

Graphene Ox 3-4 week old Swiss albino  Following IV administration graphene accumulated within the lungs, liver, spleen and kidneys. 145 mice  Confocal Raman spectroscopy demonstrated that the carboxylated graphene underwent biodegradation over 90 days in various organs. RAW 264.7 macrophages  Spleen bound graphene showed the most significant degradation after 90 days.  The presence of the material within all organs resulted in considerable toxicity Degradation of the material was also demonstrated by macrophages in vitro over a week.

Graphene GO-PEG HRP  GO showed remarkable cytotoxicity to the 3 cell lines. 102  Reduced cytotoxicity was observed for GO-BSA, rGO PEG, and rGO-BSA, and particularly for GO- GO-BSA HL-7702 PEG, compared to GO alone. rGO PEG MRC-5  Using a HRP model GO-PEG, GO-BSA, rGO-PEG and rGO-BSA were not degradable, unlike the rGO-BSA U937, extensive degradation of GO.  GO-SS-PEG was designed which showed high solubility. The structure demonstrated rapid Balb/C mice biodegradability via HRP after the cleavage of the disulphide bridge (S-S).  In vivo, GO-SS-PEG mainly accumulated in the liver and the spleen following IV administration.

SWNT Ox XO  Macrophages in vitro and in vivo up took SWNTs. 138  After 28 days following pulmonary exposure, SWNT had been majorly cleared from the lungs of wild

SIN-1 type mice, but this was not the case in gp91phox(_/_) mice.  Macrophages induced defects in SWNT as evidenced by an increase I(D)/I(G) ratio and a decrease in Wild-type C57BL6 SWNT length. C57BL/6 mice with  Two in vitro sources of NO•, namely 1) xanthine oxidase/xanthine), with NO* donors (PAPA-NONOate NADPH-oxidase- or spermine NONOate) and 2) decomposing of SIN-1 were shown to degrade SWNTs.

deficient (gp91phox(_/_)) mice

THP-1 cells

Graphene Pristine 4–5 week old Swiss albino  Following IV administration graphene-based structures accumulated mainly within lungs followed by 166 mice the spleen, liver and kidney over a period of 24 h. Carboxylated  Pristine and carboxylated graphene induced significant cellular and structural toxicity to the lungs, PEG liver, spleen, and kidney tissues.

 PEGylated graphene did not induce such toxicity.  Over 3 months, carboxylated and PEGylated graphene demonstrated spectral features consistent with biodegradation whereas pristine graphene did not.

HRP = Horse Radish peroxidase, MPO = Myeloperoxidase, PEG= Poly ethylene glycol, BSA = Bovine serum albumin, NOX2 = Subunit of nicotinamide adenine dinucleotide phosphate-oxidase, Ox = oxidised or carboxylated, rGO = Reduced GO

2.8.4. Summary of the factors that can be used to modulate CNMs biodegradation relevant to biomedicine.

In this part of the review, we have interrogated the literature to reveal systems that have been investigated to mediate the biodegradation of CNMs and have highlighted the key factors that can influence the biodegradation kinetics. We have combined this information below (Figure 2.20) in an attempt to help indicate what factors may be useful in the design of CNMs with controlled biodegradation kinetics in the contexts of more biologically relevant scenarios (cells – both in vitro and in vivo), to aid in the design of biomedical applications.

100

• Single Walls • Single Layers • Multiple Walls • Highly Oxidised • Low Oxidation • Multiple Layers • High Defects • Few Defects • Pristine • Enzyme • Intermediate • Aggregated Substrates Dispersibility • Inhibitory • Stimulatory • PEGylation Substrates Proteins • Cellular

Activation % degradationof%

Short Term Intermediate Term Long Term (Days - Weeks) (Weeks -Months) (Months - Years) Time

Figure 2.20 An illustration of the various parameters that can affect the biodegradation kinetics of different CNMs in view of biomedical applicability.

2.9. Products of degradation and their health effects.

Thus far we have discussed the degradability of CNMs as well as factors that can influence the kinetics of their degradation. However, as alluded during the course of this review, aside from the generation of CO2, various intermediate by-products or partially degraded

84 structures can be produced. Many studies have demonstrated that these by-products include but are not limited to oxidised polyaromatic hydrocarbons (o-PAHs). These structures are known carcinogenic agents.168, 169 The effect of these CNM by-products on human health is therefore of key importance when developing any CNM containing system that may biodegrade.

Using an in vitro electro-optical array, Pan et al. directly exposed DNA to the by-products of HRP mediated degradation of carboxylated SWNTs obtained between days 5 to 10.125 The results showed that the obtained by-products induced DNA damages, with the highest damages induced by SWNT degradation by-products obtained at day 7. After day 7, DNA damage was less pronounced. The authors confirmed their findings in A549 lung carcinoma cells. This study suggests that a rapid and complete degradation of carboxylated CNMs is necessary in order to reduce the presence of DNA damaging structures. Farrera et al. also tested the toxicity of carboxylated SWNTs that had undergone 18 h MPO mediated biodegradation.132 Freshly isolated human neutrophils were used in the model for the toxicity analysis. In support of the damaging effects described by Pan and co-workers125, the authors observed that there was a significant loss of cell viability after 6 h incubation with the degraded structures, compared to controls132.

In an experiment highly relevant to occupational exposures, Kagan et al. administered non- degraded, partially or fully degraded carboxylated SWNTs (0, 12 or 24 h of MPO mediated degradation, respectively) to mice via pharyngeal instillation.84 Non-degraded carboxylated SWNTs elicited an acute inflammatory response characterised by the formation of tissue granulomas. Partially degraded SWNTs did illicit some inflammatory response although not as severe as non-degraded structures. And fully degraded samples induced neither inflammation nor resulted in the formation of granulomas. This was later confirmed by the same team in which they demonstrated that the lung inflammatory response to CNT pharyngeal aspiration was greater in MPO deficient mice in comparison to MPO competent mice.162 Similarly, it has been shown that following the incubation of CNTs in Gamble’s solution (to replicate degradation in lysosomes), the subsequent administration of the degraded CNTs into the peritoneal cavity induce less inflammation and fibrosis than non- degraded samples.89

Overall, in contrast to the in vitro experiments, studies in vivo support the mitigation of toxicity via biodegradation. However, in order to truly confirm this, long term biodegradation

85 studies should be completed following exposure to CNM via different routes. The effects of in vivo produced biodegradation products could then be assessed as they are being produced. This is in line with a recent comment on the need for more long term assessments of the interaction of nanostructures with organisms.170

2.10. Biodegradation requirements for CNM based biomedical applications and beyond.

In order to efficiently demonstrate the important relevance of understanding the biodegradability of CNMs, we end this review by discussing, using case studies, the importance of considering biodegradation in the context of biomedical applications where CNMs have been applied, and how different applications demand materials that biodegrade with specific kinetics. We focus on 3 major biomedical areas where the applications of CNMs have been investigated, namely drug delivery, tissue engineering and diagnostics.

2.10.1. Drug delivery.

Following decades of clinical and preclinical studies, it is possible to contemplate the ideal attributes of a drug delivery nano-vector. The ideal system should hold a high drug payload, prevent untoward release of drug in non-target regions, preserve the physicochemical integrity of the loaded drug molecule, and should deliver the drug molecule to the desired target location with high efficiency. However, foremost the carrier must be biocompatible and biodegradable.171 There are various examples in the literature where the intended use of the CNMs as a delivery vector is simply to hold, transport and deposit a drug, or carry out a therapeutic function at a target site, such as in some cancer treatment regimens. However, more often than not, there is also some unwanted non-specific targeting and to some extent the CNM vector delivery system remains in the body after it has completed its purpose.

For example, Liu et al. designed a PEGylated SWNT based delivery system in which the active ingredient, paclitaxel, was conjugated to the PEG group via an ester bond.172 When administered intravenously, although the delivery of paclitaxel to the tumour was more specific when bound to the PEG-SWNT delivery system, there was also a noticeable uptake of materials in the spleen and liver. The authors found that the ester linkage was cleaved in these organs and the free drug was excreted, however PEG-SWNTs remained present within the tissue. Tian et al. also reported the development of an intravenously delivered, targeted GO-based system for the delivery of various anticancer agents including camptothecin.173. Similarly to Liu et al.172, there was a parallel accumulation within the organs of the MPS.

In both of these proof of concept studies, following the release of the payload drug and the successful completion of its therapeutic function at the target region, the drug delivery vectors ideally should be rapidly cleared from the body or be safely biodegraded within days to several weeks. This is in order to prevent long-term bioaccumulation at the targeted site, which could have negative consequences. This is even more critical where repeated doses are required as is often the case in some cancer therapies.174 Moreover, where the nanocarriers deposit in non-target regions such as MPS-related organs, these constructs should also be safely biodegraded or cleared.

The use of SWNTs opposed to MWNTs in Liu et al. 172 study is advantageous regarding the materials biodegradability.88 In addition, the surface chemistry should also be considered. The PEGylated SWNT system employed was based on a non-covalent interaction between a synthesised phospholipid-PEG molecule with the outer surface of pristine SWNTs. Literature studies indicate that, once the PEGylated SWNT structure is taken up by neutrophils or macrophages, the PEG chains can be initially ‘stripped’ from the surface of the SWNT prior to biodegradation of the CNM.106 The unfunctionalised pristine surface of the SWNT will then be exposed. Such unfunctionalised surfaces have been shown to be resistant to biodegradation under both in vitro 84, 153 and in vivo166 contexts. Ultimately, this means that any bodily retained SWNT are likely to be biodurable and remain within the body over the long term if they are not excreted over time.

In order to avoid this scenario, the surface of the SWNT could be covalently functionalised in order to introduce defects into the crystalline graphene structure. The literature indicates that at the cellular level, the type of surface chemistry is not as important as the degree of defects induced, although carboxylated samples tend to undergo a more rapid biodegradation initially.155 This means that materials that remain in the body following the release of their therapeutic cargo can be rapidly biodegraded. Moreover proteins that are present in vivo, such as plasma proteins, may spontaneously adsorb to the surface of the defected nanomaterial. Studies as described previously have shown that such protein coatings can act to stimulate cells such as macrophages108 and neutrophils84, 109, 132 in order to accelerate the defected materials’ biodegradation. Moreover, strategies can also be used that specifically functionalise the defected surfaces with substrates that will purposefully further predispose the CNM to biodegradation that could be mediated by specific enzymes in vivo.100 Such steps will help to promote rapid biodegradation of the vector.

In terms of the PEGylated GO-based nanovector described by Tian et al,173, the structure was shown to be 7.5 nm thick, which may be the result of few stacked layers of functionalised GO. Furthermore, GO by nature is a highly defected material that is decorated with various covalently attached oxygen functionalities. According to literature research, this structure is particularly susceptible to degradation.59 However, the attached PEG chains may initially delay the biodegradation of the CNM.166 To increase the kinetics, strategies can be used that facilitate the controlled release of the PEG chains. For example by functionalising the PEG chain to the GO using disulphide bridges, these bonds could cleave intracellularly102 as a result of being reduced by intracellularly present reducing species such as glutathione. This will result in the release of the highly defected GO that is expected to display more rapid biodegradation kinetics.

In other cases, where a controlled and sustained drug release is desired175, the biodegradation of the CNM-based systems should be more delayed and suit the kinetics of drug release. GO has been investigated in drug delivery systems to enhance and prolong the loading of drugs due to its high specific surface area that can be used to non-covalently adsorb drug molecules. This has been demonstrated by using GO to enhance the loading and to delay the release of the anti-cancer drug doxorubicin from a hyaluronic acid hydrogel- GO composite formulation. The nanocomposite was shown via in vitro testing to behave as an anticancer drug reservoir and release the drug over a 10 day period.176 Hong et al. also designed a sustained protein antigen release system for ovalbumin (OVA) based on a layer by layer assembly of hydrolytically degradable multilayer films using a poly (β-amino ester) (POLY1) with attached OVA.177 Within this system, GO sheets were used to cap each of the individual layers of POLY1. The GO layers were found to modulate the original burst release kinetics of the OVA from the POLY1 layers. When GO was incorporated into the system, the greater the number of layers the slower the release of OVA. This was to the extent that a 20 bilayer film system resulted in a linear release of OVA over 90 days. The authors also went on to show how sequential protein release could be achieved by incorporating different fluorescently tagged OVA moieties into different compartments of the same multilayer structure.177

In these cases, where sustained and/or controlled release kinetics of one or more active ingredients is desired, the biodegradation kinetics of the CNMs (GO in these examples) should be tailored such that the nanomaterials along with their drug payload are not prematurely broken down or cleared. Otherwise, this would jeopardise the potential

89 therapeutic strategy or even worse result in a burst release and thus an overdose. According to the literature, this can be achieved by using multi-layered materials59 and or materials that are more pristine.101, 123 In the layer-by-layer assemblies described above, this could be employed without the further use of surfactant or PEG chains to preserve the dispersibility of the system. This is because these particular delivery systems are to be implanted and not intravenously administered. However, such strategies would not be possible for intravenously administered systems due to the risk of agglomeration and poor colloidal stability, which would result in significant accumulation of the materials within narrow diameter capillaries.47, 55 In these cases, structures could be attached to the surface of the material including PEG chain102 or enzyme inhibitory substrates100 that would delay biodegradation yet preserve the dispersibility of the materials.

In all cases, however, care should be taken not to increase the toxicity of the material as a result of the modification of its physicochemical properties. For example, the type and degree functionalisation of the surfaces of CNMs can alter their biocompatibility, with more highly functionalised materials generally being more biocompatible.178 Changes in the surface chemistry can also effectively change the biocompatibility of the material by altering the combination of proteins that adsorb to its surface (termed the protein corona) in vivo. This was demonstrated in a study investigating the different biocompatibilities of functionalised GO derivatives.139 Here the functionalisation of GO with poly acrylic acid or poly ethylene glycol altered the materials biocompatibility due to variations in the abundances of adsorbed IgG. At the same time, such phenomena could be used in strategies to engineer the material’s biodegradability profile to promote biodegradation. It has been shown that certain adsorbed proteins, including IgG, can stimulate certain phagocytic cells and increase their biodegradability capacity84, 109, 132. At the same time, too much stimulation can be harmful179, 180 and thus compromises should be met that allow suitable biodegradation kinetics but avoid harmful toxicological consequences.

In addition to the specific chemistries of the resultant materials, the means by which they are derived should also be considered. Studies have shown that rGO formed via the hydrazine mediated reduction of GO, is more toxic than rGO derived through ascorbic acid mediated reduction of GO.181

It can be seen that when CNMs structures are being altered in order to tailor their biodegradability, studies should be performed to ensure that their biocompatibility profiles

90 are maintained. Furthermore, the toxicological profile of the breakdown products should also be considered, although the limited literature available thus far indicate that biodegraded CNMs demonstrate more favourable in vivo biocompatibility profiles compared to their non- degraded counterparts.84, 89 Despite this, in vitro studies appear to suggest the contrary125, 132 therefore particular attention should be paid to this aspect. Moreover, as the chemistry of the CNMs changes so will their biodegradation by-products.

2.10.2. Tissue engineering.

In addition to drug delivery, CNMs have been explored for their utility in tissue engineering. Tissue engineering is described as ‘an interdisciplinary field that applies the principles of engineering and the life sciences toward the development of biological substitutes that restore, maintain, or improve tissue function’.182 The field in general heavily depends on the engineering of scaffolds that serve as templates that help to promote the regeneration of damaged or lost tissues by temporarily mimicking the native extracellular matrix.183 CNMs have been explored for the design of novel scaffolds. Some key specifications should be considered when designing a tissue engineering scaffold, regardless of the material used. The scaffold should be biocompatible and enable the healthy growth and proper functioning of cells. Moreover, the scaffold must not induce immunological consequences that could for example result in rejection or anaphylaxis. It should have mechanical properties and physical features that are consistent with the tissue to be regenerated and this can differ quite dramatically between tissues, for example bone184 vs. cardiac tissue.185 The scaffold should also have an architecture that includes intersected cavities of defined dimensions that allow the movement of cells as well as the removal of waste and the provision of nutrients and oxygen, that are necessary to support a healthy regeneration.183 Furthermore, the scaffold should be consistent with any specialised properties of the tissue to be regenerated such as allow the conductance of electrical signals in the case of neuronal tissue.186

Besides all of these essential characteristics, the selection of biodegradable or non- biodegradable materials is connected to the end application of the scaffold.183 In both cases, biodegradability of the scaffold is a crucial parameter that needs to be assessed, especially if biodegradation could lead to negative impacts on the tissue to be replaced or regenerated. In addition, aside from the toxicological implications, the mechanical properties or any other particular assets of the scaffold could be affected by non-desired or uncontrolled biodegradation75, 154 and thereby compromise the ability of the scaffold to appropriately

91 support regeneration. Biodegradation studies are hence essential and must be undertaken. For biodegradable scaffolds under ideal circumstances, as cells migrate or populate the scaffold occupied region and the innate biological matrix is gradually re-established; there should be a coincident controlled removal of the implanted scaffold via biodegradation. The by-products of biodegradation must be non-toxic at the doses that they are produced, and should not bio-accumulate but be cleared.187 It should be considered that the speed at which the endogenous tissue matrix is replaced in vivo will depend on the tissue type in question. Each tissue will regenerate at a different rate and this may slightly differ in a case by case manner. The kinetics of the biodegradation of the scaffold must therefore be suitably tailored.188

Although less common, some medical technologies employ non-biodegradable materials for tissue engineering.189 For example, non-biodegradable materials have been used for dental tissue engineering in the replacement of large bony defects as well as in the design of vascular grafts.190 In these cases, the materials act as constant structural supports for tissues while at the same time allow some integration of the host cells. Generally such materials are only used when complete healing or assisted regeneration is not possible. In these cases, the material must be engineered to avoid biodegradation and loss of its mechanical properties.191 Where complete healing is possible however, biodegradable materials are more attractive and advantageous since surgery is not required to remove them at a later point in time. Therefore, when designing scaffolds with CNMs, the biodegradation should be considered with regards to the aims of the tissue engineering strategy, and the type of CNMs to be employed in the designs should be chosen accordingly.

Mata et al. showed how changes in the component CNMs properties could affect the biodegradability of the tissue engineering scaffold. The authors studied the feasibility of using MWNTs in bone tissue engineering.154 The biocompatibility and biodegradability of membrane scaffolds composed of either Diels–Alder functionalised or pristine MWNTs were interrogated. The functionalised and non-functionalised MWNT membranes were subcutaneously implanted in rats and both their biocompatibility and biodegradability were assessed over 49 days. After 7 days, both membranes had elicited mild inflammatory responses that were characterised by granulomatous tissue and inflammatory infiltrates, but without associated necrosis or presence of cell debris. As for biodegradability, signs of membrane breakdown were clear for the functionalised but not the pristine membranes,

92 which maintained their structural integrity. When the cells of the inflammatory infiltrate (i.e. cells immediately surrounding the biodegrading functionalised membrane) were closely inspected, some appeared to contain small CNT-like fibrillary structures, indicating that the breakdown of the scaffold may have been mediated by these infiltrating cells. This study clearly illustrated the need for biodegradation studies to be conducted when developing scaffolds made of CNMs. It also highlights the need to explore how the biodegradation of scaffolds can be modulated by chemical functionalisation to enhance or retard the biodegradation kinetics in accordance with the targeted application.

Where such scaffolds are used as therapeutic platforms to support the regeneration of more sensitive biological systems, such as the central nervous system (CNS), the need to assess the biodegradation kinetics of the composing materials becomes more critical. In this case if the materials were to biodegrade, the kinetics of this process must be carefully tuned with the CNS repair kinetics. To give an idea about the rate of neural regeneration, Ellis-Behnke et al. used a biodegradable peptide nanofiber based scaffold to support the healing of a 1.5 mm deep and 2.0 mm wide wound in the midbrain anatomical region.192 With the aid of this scaffold, the wound took 30 days to close, allowing both structural and functional recovery of the neuronal network.

The design of CNM tissue engineering based scaffolds with controllable biodegradation kinetics will evidently be difficult to engineer and require significant investigation. This is because the rate of biodegradation, as well as the material’s mechanical properties, must be carefully adapted to the tissue in question. Based on our review, in order to increase the biodegradation kinetics of CNMs, researchers could employ thinner59, 88 and highly defected materials84, 101 that are expected to promote biodegradation, or employ the contrary where more prolonged biodegradation kinetics are desired. Further strategies as discussed in Section 2.10.1 could also be employed. However, irrespective of the strategy selected, attention must be paid to not change the materials so much that they are no longer suitable to be used as scaffold materials. Studies have demonstrated that chemical functionalisation of the surface of CNMs can alter functional properties. For example, it has been shown that by varying the functionalisation chemistries of CNMs, the mechanical properties75 as well as the electrical properties193 can be sensitively affected. Such changes could disturb the materials suitability as a tissue engineering scaffold.

Alongside biodegradation, the by-products’ chemical nature and rate of production must be determined here too. A build up for instance of a noxious by-product could be devastating.125 This is even more relevant for scaffolds, compared to drug delivery, as the amount of material contained within tissue engineering scaffold is generally quite considerable. For example, in the case of Mata et al., each animal was implanted with 4 mg of SWNTs.154 However, when by-products obtained from enzyme derived CNM biodegradation in vitro were placed in vivo, they were more biocompatible compared to the starting materials. Thus far, only limited exposed bodily systems have been investigated and so more research is needed to expand this knowledge. All these key considerations and associated questions are supporting the call for a bigger research effort on biodegradability of CNMs in relation to their therapeutic applications, especially when considering them in invasive biomedical contexts.

Somewhat related to the application of CNMs in tissue engineering, where the materials are used in the design and fabrication of prosthetics or artificial body parts such as skin194, it is extremely important that the structure of the materials remain intact. Thus, degradation should be avoided entirely since the design is generally expected to last as long as possible. Therefore, every effort should be taken to ensure this by using pristine101 or thick (and multiwalled) structures59, 88 that have been shown in the literature to be resistant to biodegradation, or indeed engineering efforts can be made to isolate them completely from biological and chemical influences if possible.

2.10.3. Diagnostics and monitoring.

CNMs and their derivatives have also been explored for their applicability in the design of imaging probes, including targeted contrast agents, for the purposes of diagnosis and monitoring. The purpose of such probes is to facilitate the characterisation and measurement of biological processes. In the case of molecular probes, the resolution of such imaging should be to the cellular, subcellular and if possible to the molecular levels. Imaging probes have provided much benefit in the diagnosis and management of medical conditions such as cancer, neurological and cardiovascular diseases. They usually consist of several elements including a signalling or contrasting agent and a targeting moiety connected by a linker group.195 The unique properties of CNMs are expected to enhance this field, enabling greater diagnostic and imaging capabilities. When designing imaging probes, several ideal properties should be considered. The imaging probe should first of all be

94 biocompatible at the doses used. It should bind with high affinity and selectivity to the target site, for example a cancer receptor or antigen. It should enable highly sensitive detection, allowing the characterisation of small targets without requiring a large exposure in order to obtain meaningful results. The probe should also display a high in vivo stability until the imaging or measurements are completed. In addition to these specifications, after the probes are no longer required they should undergo rapid and safe clearance and/or complete biodegradation.195 The same strategies discussed in Sections 2.10.1 and 2.10.2 could also be applied here as well, however the biodistribution and precision of the biomarker should be importantly maintained.

In accordance with this, when CNMs are incorporated into the design of these specialised pharmaceutical agents, and if clearance is incomplete, CNM-based probes should biodegrade. In contrast, where clearance is complete, biodegradation will not be an essential design criterion. Delinger et al. developed a CD36 targeted liposome – gadolinium metallofullerene nano-hybrid for the detection of CD36-expressing foam cells.196 These cells are associated with atherosclerotic plaques and contribute to plaque dislodgment resulting in the formation of emboli. In vivo studies proved that the hybrid systems were not retained in the body over time. Within 7 days, approximately 85 % of the nano systems were excreted. The remaining materials were detected within the kidneys and liver. In such cases, where there is minimal in vivo retention of a biocompatible imaging or contrast agent, biodegradability is not a major design requirement. However, assessing the potential biodegradation of the remaining materials could still be essential to understand whether or not toxic by-products are formed. In the presented case for example196, if the carbon truncated icosahedron cage of the gadolinium metallofullerenes were to be biodegraded, aside from the unknown carbon-based biodegradation products, toxic gadolinium ions could be released.197

In more complex ‘smart’ designs composed of multiple cleavable entities, more attention must be paid to the in vivo fate of each individual component. For example, Yue et al. described the design of a GO-based inducible probe for early tumour diagnosis, in which a fluorophore (Cy5) was conjugated to a GO nanosheet via a short peptide linker.198 While attached to the GO nanosheet, the fluorescence of the fluorophore was quenched due to fluorescence resonance energy transfer. When the probe entered the vicinity of the tumour, tumour specific enzymes cleaved the peptide linker, thus liberating the fluorophore from the GO. The detectable fluorescence of the fluorophore was thereby restored. The probe’s efficacy was demonstrated

95 both in vitro and in vivo following intravenous administration into a mouse model bearing a subcutaneous Lewis lung carcinoma cell derived tumour. However, in vivo fate of the released GO nano sheet with the associated cleaved peptide fragment was not further determined. Ideally the long term fate (clearance and/ or biodegradation) and biocompatibility of the functionalised GO flakes should be assessed. In this instance, following the cleavage of the peptide linker, the GO flakes should undergo rapid clearance and or biodegradation.

Where CNMs are used as key part of in vivo devices for the monitoring and detection of various 199 molecules, such as H2O2 , the materials should not undergo biodegradation in the presence of molecules of the biological media (e.g. H2O2), unless this is of course a specific requirement for the biosensing mechanism. Therefore in these cases, biodegradability analysis should be undertaken to ensure the robustness and reusability of the in vivo biosensor. Ideally, the CNM component should be tailored to be resistant to biodegradation using strategies discussed in previous sections.

2.10.4. Summary of the need to consider CNM biodegradation in relation to their biomedical applications.

It can be seen from Section 2.10 that the biodegradability of CNMs and the kinetics of these processes must be carefully considered and tuned to enable the CNMs to be effectively and successfully applied to biomedical technologies. It is both pertinent with regards to the toxicological implications of these unique materials but also for the correct functioning of the biomedical strategies. Below we summarise the varying biodegradability kinetic requirements of CNMs when incorporated within different biomedical strategies (Figure 2.21).

Drug delivery

Imaging Rapid Controlled Drug Delivery

Tissue Engineering

Electronic devices

Prosthetics

Degradation Kinetics Slow

Short Term Intermediate Term Long Term (Days - Weeks) (Weeks - Months) (Months – Years +) Time until complete degradation

Figure 2.21 An illustration of the different biodegradation kinetic requirements of CNMs when incorporated into various biomedical strategies.

2.11. Conclusion.

The recent scientific and technological interest into CNMs, particularly graphene, has been rapidly intensifying as their applicability to enhance various fields becomes increasingly apparent. However, their durability was perceived as a potential drawback for some applications. Indeed for a significant period of time, CNMs were believed to be highly durable due to the high chemical stability of their graphitic structure. This belief gave the impression that the wide usage potential of CNMs may never be realised, particularly in biomedicine where biodegradability could be a prerequisite criteria in most cases. It was also considered as a cause for concern for ecosystems if CNMs were to be unintentionally released to the environment. Recent reports have however demonstrated that CNMs can undergo biodegradative processes under defined conditions, and that these processes may be realised within living systems.

In this review, we conducted an analysis of the factors that can affect the biodegradation kinetics under the influence of different systems such as oxidative chemicals, enzymes, or cells (in vivo and in vitro). We find that factors such as the number of layers (or walls), the oxidation state, intrinsic defects, and the dispersibility as well as nanomaterial coatings can be tailored to optimise the biodegradation kinetics. We have found that, as biodegradative systems get more complicated, going from isolated chemicals to living organisms, the responses of the system to the presence of nanomaterials and their design (either due to purposeful engineering or unforeseen factors) can affect the biodegradation kinetics, in ways not predicted via the use of non-living systems. In order to demonstrate the requirement of such understanding, we highlight the need for application-specific biodegradation kinetics with respect to three different biomedical areas. In general, there is a need for investigations to determine the biodegradability of CNMs, such as GO, in vivo over the long term and administered via different routes. Moreover, future investigations should attempt to further understand the biodegradability of the CNMs when incorporated within biomedical strategies to further elaborate on the design of applications with programmed and tuned biodegradability.

2.12. References.

1. F. Banhart, Beilstein J. Nanotechnol. 2015, 6, 559-569. Chains of carbon atoms: A vision or a new nanomaterial? 2. A. Hirsch, Nat. Mater., 2010, 9, 868-871. The era of carbon allotropes. 3. E. A. Ekimov, V. A. Sidorov, E. D. Bauer, N. N. Mel'nik, N. J. Curro, J. D. Thompson and S. M. Stishov, Nature, 2004, 428, 542-545. Superconductivity in diamond. 4. L. Pauling, Proc. Natl. Acad. Sci. U. S. A., 1966, 56, 1646-1652. The structure and properties of graphite and boron nitride. 5. H. W. Kroto, J. R. Heath, S. C. O'Brien, R. F. Curl and R. E. Smalley, Nature, 1985, 318, 162-163. C60: Buckminsterfullerene 6. S. Iijima, Nature, 1991, 354, 56-58. Helical microtubules of graphitic carbon 7. S. Iijima and T. Ichihashi, Nature, 1993, 363, 603-605. Single-shell carbon nanotubes of 1-nm diameter 8. S. Iijima, M. Yudasaka, R. Yamada, S. Bandow, K. Suenaga, F. Kokai and K. Takahashi, Chem. Phys. Lett., 1999, 309, 165-170. Nano-aggregates of single- walled graphitic carbon nano-horns 9. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666-669. Electric field effect in atomically thin carbon films. 10. N. Arora and N. N. Sharma, Diamond Relat. Mater., 2014, 50, 135-150. Arc discharge synthesis of carbon nanotubes: comprehensive review. 11. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater. (Weinheim, Ger.), 2010, 22, 3906-3924. Graphene and graphene oxide: synthesis, properties, and applications. 12. A. A. Popov, S. Yang and L. Dunsch, Chem. Rev. (Washington, DC, U. S.), 2013, 113, 5989-6113. 13. A. K. Geim and K. S. Novoselov, Nat. Mater., 2007, 6, 183-191. The rise of graphene. 14. M. F. De Volder, S. H. Tawfick, R. H. Baughman and A. J. Hart, Science, 2013, 339, 535-539. Carbon nanotubes: present and future commercial applications. 15. K. Kostarelos and K. S. Novoselov, Nat. Nanotechnol., 2014, 9, 744-745. Graphene devices for life. 16. M. Peplow, Nature, 2015, 522, 268-269. Graphene booms in factories but lacks a killer app.

17. A. Helland, P. Wick, A. Koehler, K. Schmid and C. Som, Environ. Health Perspect., 2007, 115, 1125-1131. Reviewing the environmental and human health knowledge base of carbon nanotubes. 18. K. Bhattacharya, S. P. Mukherjee, A. Gallud, S. C. Burkert, S. Bistarelli, S. Bellucci, M. Bottini, A. Star and B. Fadeel, Nanomedicine: Nanotechnology, Biology and Medicine, 2016, 12, 333-351. Biological interactions of carbon-based nanomaterials: From coronation to degradation. 19. C. Bussy, A. Bianco, M. Prato and K. Kostarelos, in Handbook of Safety Assessment of Nanomaterials - From Toxicological Testing to Personalized Medicine, ed. B. Fadeel, Stanford Publishing Pte Ltd Editon edn., 2014. 20. C. Bussy, H. Ali-Boucetta and K. Kostarelos, Acc. Chem. Res., 2013, 46, 692-701. Safety considerations for graphene: lessons learnt from carbon nanotubes. 21. A. M. Pinto, I. C. Gonçalves and F. D. Magalhães, Colloids Surf., B, 2013, 111, 188- 202. Graphene-based materials biocompatibility: a review. 22. A. Montagner, S. Bosi, E. Tenori, M. Bidussi, A. A. Alshatwi, M. Tretiach, M. Prato and Z. Syrgiannis, 2D Mater., 2017, 4, 012001. Ecotoxicological effects of graphene- based materials. 23. Z. Chen, J. I. Wu, C. Corminboeuf, J. Bohmann, X. Lu, A. Hirsch and P. Schleyer, Phys. Chem. Chem. Phys., 2012, 14, 14886-14891. Is C60 buckminsterfullerene aromatic? 24. R. Taylor, C. R. Chim., 2006, 9, 982-1000. Addition reactions of fullerenes. 25. R. Bakry, R. M. Vallant, M. Najam-ul-Haq, M. Rainer, Z. Szabo, C. W. Huck and G. K. Bonn, Int. J. Nanomed., 2007, 2, 639-649. Medicinal applications of fullerenes. 26. J. B. Howard, J. T. McKinnon, Y. Makarovsky, A. L. Lafleur and M. E. Johnson, Nature, 1991, 352, 139-141. Carbon shells in flames. 27. A. Chuvilin, U. Kaiser, E. Bichoutskaia, N. A. Besley and A. N. Khlobystov, Nat. Chem., 2010, 2, 450-453. Direct transformation of graphene to fullerene. 28. A. Eatemadi, H. Daraee, H. Karimkhanloo, M. Kouhi, N. Zarghami, A. Akbarzadeh, M. Abasi, Y. Hanifehpour and S. W. Joo, Nanoscale Res. Lett., 2014, 9, 393-393. Carbon nanotubes: properties, synthesis, purification, and medical applications. 29. A. Aqel, K. M. M. A. El-Nour, R. A. A. Ammar and A. Al-Warthan, Arabian J. Chem., 2012, 5, 1-23. Carbon nanotubes, science and technology part (I) structure, synthesis and characterisation. 30. M. Yudasaka and S. Iijima, in Carbon Nanotubes and Related Structures, Wiley-VCH Verlag GmbH & Co. KGaA, Editon edn., 2010, pp. 385-404. 31. Y. Tahara, M. Nakamura, M. Yang, M. Zhang, S. Iijima and M. Yudasaka, Biomaterials, 2012, 33, 2762-2769. Lysosomal membrane destabilization induced by

100

high accumulation of single-walled carbon nanohorns in murine macrophage RAW 264.7. 32. N. Karousis, I. Suarez-Martinez, C. P. Ewels and N. Tagmatarchis, Chem. Rev. (Washington, DC, U. S.), 2016, 116, 4850-4883. Structure, properties, functionalization, and applications of carbon nanohorns. 33. C. Pagura, S. Barison, C. Mortalò, N. Comisso and M. Schiavon, Nanosci. Nanotech. Lett., 2012, 4, 160-164. Large scale and low cost production of pristine and oxidized single wall carbon nanohorns as material for hydrogen storage. 34. P. J. F. Harris, S. C. Tsang, J. B. Claridge and M. L. H. Green, J. Chem. Soc., Faraday Trans., 1994, 90, 2799-2802. High-resolution electron microscopy studies of a microporous carbon produced by arc-evaporation. 35. M. Yudasaka, S. Iijima and V. H. Crespi, in Carbon Nanotubes: Advanced Topics in the Synthesis, Structure, Properties and Applications, eds. A. Jorio, G. Dresselhaus and M. S. Dresselhaus, Springer Berlin Heidelberg, Berlin, Heidelberg, Editon edn., 2008, pp. 605-629. 36. T. Kawai, Y. Miyamoto, O. Sugino and Y. Koga, Physical Review B, 2002, 66, 033404. Nanotube and nanohorn nucleation from graphitic patches: Tight-binding molecular dynamics simulation. 37. F. Kokai, A. Koshio, D. Kasuya, K. Hirahara, K. Takahashi, A. Nakayama, M. Ishihara, Y. Koga and S. Iijima, Carbon, 2004, 42, 2515-2520. Three nanostructured graphitic particles and their growth mechanisms from high-temperature carbon vapor confined by Ar gas. 38. K. Kostarelos and K. S. Novoselov, Science, 2014, 344, 261-263. Materials science. Exploring the interface of graphene and biology. 39. R. Peierls, Ann. Inst. Henri Poincare, 1935, 5. Quelques propriétés typiques des corps solides. 40. N. D. Mermin, Phys. Rev., 1968, 176, 250-254. Crystalline order in two dimensions. 41. K. S. Novoselov, Z. Jiang, Y. Zhang, S. V. Morozov, H. L. Stormer, U. Zeitler, J. C. Maan, G. S. Boebinger, P. Kim and A. K. Geim, Science, 2007, 315, 1379-1379. Room-temperature quantum Hall effect in graphene. 42. M. Prato, K. Kostarelos and A. Bianco, Acc. Chem. Res., 2008, 41, 60-68. Functionalized carbon nanotubes in drug design and discovery. 43. M. Zhang, M. Yudasaka, K. Ajima, J. Miyawaki and S. Iijima, ACS Nano, 2007, 1, 265-272. Light-assisted oxidation of single-wall carbon nanohorns for abundant creation of oxygenated groups that enable chemical modifications with proteins to enhance biocompatibility.

101

44. H. Ali-Boucetta, D. Bitounis, R. Raveendran-Nair, A. Servant, J. Van den Bossche and K. Kostarelos, Adv. Healthcare Mater. 2013, 2, 433-441. Purified graphene oxide dispersions lack in vitro cytotoxicity and in vivo pathogenicity. 45. B. Wang, B. Luo, M. Liang, A. Wang, J. Wang, Y. Fang, Y. Chang and L. Zhi, Nanoscale, 2011, 3, 5059-5066. Chemical amination of graphene oxides and their extraordinary properties in the detection of lead ions. 46. Z. Liu, J. T. Robinson, X. Sun and H. Dai, J. Am. Chem. Soc., 2008, 130, 10876- 10877. PEGylated nanographene oxide for delivery of water-insoluble cancer drugs. 47. L. Illum, S. S. Davis, C. G. Wilson, N. W. Thomas, M. Frier and J. G. Hardy, Int. J. Pharm., 1982, 12, 135-146. Blood clearance and organ deposition of intravenously administered colloidal particles. The effects of particle size, nature and shape. 48. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228-240. The chemistry of graphene oxide. 49. I. A. Vacchi, C. Spinato, J. Raya, A. Bianco and C. Menard-Moyon, Nanoscale, 2016, 8, 13714-13721. Chemical reactivity of graphene oxide towards amines elucidated by solid-state NMR. 50. B. C. Brodie, Philos.Trans. R. Soc. London., 1859, 149, 249-259. On the atomic weight of graphite. 51. L. Staudenmaier, Ber. Dtsch. Chem. Ges., 1898, 31, 1481-1487. Verfahren zur darstellung der graphitsäure. 52. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339-1339. Preparation of graphitic oxide 53. D. Jasim, A., N. Lozano and K. Kostarelos, 2D Mater., 2016, 3, 014006. Synthesis of few-layered, high-purity graphene oxide sheets from different graphite sources for biology. 54. A. Bianco, H.-M. Cheng, T. Enoki, Y. Gogotsi, R. H. Hurt, N. Koratkar, T. Kyotani, M. Monthioux, C. R. Park, J. M. D. Tascon and J. Zhang, Carbon, 2013, 65, 1-6. All in the graphene family – A recommended nomenclature for two-dimensional carbon materials 55. C. Bussy, D. Jasim, N. Lozano, D. Terry and K. Kostarelos, Nanoscale, 2015, 7, 6432-6435. The current graphene safety landscape – a literature mining exercise. 56. P. Wick, A. E. Louw-Gaume, M. Kucki, H. F. Krug, K. Kostarelos, B. Fadeel, K. A. Dawson, A. Salvati, E. Vázquez, L. Ballerini, M. Tretiach, F. Benfenati, E. Flahaut, L. Gauthier, M. Prato and A. Bianco, Angew. Chem., Int. Ed., 2014, 53, 7714-7718. Classification framework for graphene-based materials.

102

57. D. A. Jasim, C. Menard-Moyon, D. Begin, A. Bianco and K. Kostarelos, Chem. Sci., 2015, 6, 3952-3964. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. 58. D. A. Jasim, H. Boutin, M. Fairclough, C. Ménard-Moyon, C. Prenant, A. Bianco and K. Kostarelos, Applied Materials Today, 2016, 4, 24-30. Thickness of functionalized graphene oxide sheets plays critical role in tissue accumulation and urinary excretion: A pilot PET/CT study. 59. R. Kurapati, J. Russier, M. A. Squillaci, E. Treossi, C. Ménard-Moyon, A. E. Del Rio- Castillo, E. Vazquez, P. Samorì, V. Palermo and A. Bianco, Small, 2015, 11, 3985- 3994. Dispersibility-dependent biodegradation of graphene oxide by myeloperoxidase. 60. S. Marchesan, K. Kostarelos, A. Bianco and M. Prato, Mater. Today, 2015, 18, 12- 19. The winding road for carbon nanotubes in nanomedicine. 61. D. Bitounis, H. Ali-Boucetta, B. H. Hong, D. H. Min and K. Kostarelos, Adv. Mater., 2013, 25, 2258-2268. Prospects and challenges of graphene in biomedical applications. 62. S. Parasuraman, J. Pharmacol. Pharmacother., 2011, 2, 74-79. Toxicological screening. 63. G. P. Kotchey, S. A. Hasan, A. A. Kapralov, S. H. Ha, K. Kim, A. A. Shvedova, V. E. Kagan and A. Star, Acc. Chem. Res., 2012, 45, 1770-1781. A natural vanishing act: the enzyme-catalyzed degradation of carbon nanomaterials. 64. M. Chen, X. Qin and G. Zeng, Trends Biotechnol., 2017, 35, 836-846. Biodegradation of Carbon Nanotubes, Graphene, and Their Derivatives. 65. W.-B. Ko and K.-N. Baek, Phys. Solid State, 2002, 44, 424-426. The oxidation of fullerenes (C60, C70) with various oxidants under ultrasonication. 66. A. L. Balch, D. A. Costa, B. C. Noll and M. M. Olmstead, J. Am. Chem. Soc., 1995, 117, 8926-8932. Oxidation of buckminsterfullerene with m-chloroperoxybenzoic acid. characterization of a Cs isomer of the diepoxide C60O2. 67. H. Yu, B. Zhang, C. Bulin, R. Li and R. Xing, Sci. Rep., 2016, 6, 36143. High-efficient synthesis of graphene oxide based on improved Hummers method. 68. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806-4814. Improved synthesis of graphene oxide. 69. C. Fan, W. Li, X. Li, S. Zhao, L. Zhang, Y. Mo and R. Cheng, Chin. Sci. Bull., 2007, 52, 2054-2062. Efficient photo-assisted Fenton oxidation treatment of multi-walled carbon nanotubes.

103

70. L. Jiao, L. Zhang, X. Wang, G. Diankov and H. Dai, Nature, 2009, 458, 877-880. Narrow graphene nanoribbons from carbon nanotubes. 71. D. V. Kosynkin, A. L. Higginbotham, A. Sinitskii, J. R. Lomeda, A. Dimiev, B. K. Price and J. M. Tour, Nature, 2009, 458, 872-876. Longitudinal unzipping of carbon nanotubes to form graphene nanoribbons. 72. M. Abdolkarimi-Mahabadi and M. Manteghian, Fullerenes, Nanotubes, Carbon Nanostruct., 2015, 23, 860-864. Chemical oxidation of multi-Walled carbon nanotube by sodium hypochlorite for production of graphene oxide nanosheets. 73. X. Li, X. Wang, L. Zhang, S. Lee and H. Dai, Science, 2008, 319, 1229-1232. Chemically derived, ultrasmooth graphene nanoribbon semiconductors. 74. W. Xing, G. Lalwani, I. Rusakova and B. Sitharaman, Part. Part. Syst. Charact., 2014, 31, 745-750. Degradation of graphene by hydrogen peroxide. 75. Q. Peng, L. Han, J. Lian, X. Wen, S. Liu, Z. Chen, N. Koratkar and S. De, Phys. Chem. Chem. Phys., 2015, 17, 19484-19490. Mechanical degradation of graphene by epoxidation: insights from first-principles calculations. 76. Y. Feng, K. Lu, L. Mao, X. Guo, S. Gao and E. J. Petersen, Water Res., 2015, 84, 49-57. Degradation of (14)C-labeled few layer graphene via Fenton reaction: Reaction rates, characterization of reaction products, and potential ecological effects. 77. R. V. Lloyd, P. M. Hanna and R. P. Mason, Free Radical Biol. Med., 1997, 22, 885- 888. The origin of the hydroxyl radical oxygen in the Fenton reaction. 78. H. Bai, W. Jiang, G. P. Kotchey, W. A. Saidi, B. J. Bythell, J. M. Jarvis, A. G. Marshall, R. A. Robinson and A. Star, J. Phys. Chem. C Nanomater. Interfaces, 2014, 118, 10519-10529. Insight into the mechanism of graphene oxide degradation via the photo-Fenton reaction. 79. T. Li, C. Z. Zhang and C. Gu, Environ. Technol., 2017, 38, 1-8. Study on degrading graphene oxide in wastewater under different conditions for developing an efficient and economical degradation method. 80. J. D. Lanphere, B. Rogers, C. Luth, C. H. Bolster and S. L. Walker, Environ. Eng. Sci., 2014, 31, 350-359. Stability and transport of graphene oxide nanoparticles in groundwater and surface water. 81. W.-C. Hou, W. M. Henderson, I. Chowdhury, D. G. Goodwin, X. Chang, S. Martin, D. H. Fairbrother, D. Bouchard and R. G. Zepp, Carbon, 2016, 110, 426-437. The contribution of indirect photolysis to the degradation of graphene oxide in sunlight . 82. W.-C. Hou, I. Chowdhury, D. G. Goodwin, W. M. Henderson, D. H. Fairbrother, D. Bouchard and R. G. Zepp, Environ. Sci. Technol., 2015, 49, 3435-3443. Photochemical transformation of graphene oxide in sunlight .

104

83. E. V. Litasova, V. V. Iljin, A. V. Sokolov, V. B. Vasilyev, M. A. Dumpis and L. B. Piotrovskiy, Dokl. Biochem. Biophys., 2016, 471, 417-420. The biodegradation of fullerene C60 by myeloperoxidase. 84. V. E. Kagan, N. V. Konduru, W. Feng, B. L. Allen, J. Conroy, Y. Volkov, I. I. Vlasova, N. A. Belikova, N. Yanamala, A. Kapralov, Y. Y. Tyurina, J. Shi, E. R. Kisin, A. R. Murray, J. Franks, D. Stolz, P. Gou, J. Klein-Seetharaman, B. Fadeel, A. Star and A. A. Shvedova, Nat. Nanotechnol., 2010, 5, 354-359. Carbon nanotubes degraded by neutrophil myeloperoxidase induce less pulmonary inflammation. 85. Vlasova, II, A. V. Sokolov, A. V. Chekanov, V. A. Kostevich and V. B. Vasil'ev, Bioorg. Khim., 2011, 37, 510-521. Myeloperoxidase-induced biodegradation of single-walled carbon nanotubes is mediated by hypochlorite. 86. N. Lu, J. Li, R. Tian and Y.-Y. Peng, Chem. Res. Toxicol., 2014, 27, 1070-1077. 87. X. Liu, R. H. Hurt and A. B. Kane, Carbon N Y, 2010, 48, 1961-1969. Binding of human serum albumin to single-walled carbon nanotubes activated neutrophils to increase production of hypochlorous acid, the oxidant capable of degrading nanotubes. 88. J. Russier, C. Menard-Moyon, E. Venturelli, E. Gravel, G. Marcolongo, M. Meneghetti, E. Doris and A. Bianco, Nanoscale, 2011, 3, 893-896. Oxidative biodegradation of single- and multi-walled carbon nanotubes. 89. M. J. Osmond-McLeod, C. A. Poland, F. Murphy, L. Waddington, H. Morris, S. C. Hawkins, S. Clark, R. Aitken, M. J. McCall and K. Donaldson, Part. Fibre Toxicol., 2011, 8, 15. Durability and inflammogenic impact of carbon nanotubes compared with asbestos fibres. 90. C. A. Poland, R. Duffin, I. Kinloch, A. Maynard, W. A. H. Wallace, A. Seaton, V. Stone, S. Brown, W. MacNee and K. Donaldson, Nat. Nanotechnol., 2008, 3, 423- 428. Carbon nanotubes introduced into the abdominal cavity of mice show asbestos- like pathogenicity in a pilot study. 91. C. W. Lam, J. T. James, R. McCluskey and R. L. Hunter, Toxicol. Sci. 2004, 77, 126- 134. Pulmonary toxicity of single-wall carbon nanotubes in mice 7 and 90 days after intratracheal instillation. 92. K. Kostarelos, Nat Biotech, 2008, 26, 774-776. The long and short of carbon nanotube toxicity. 93. L. Flohe and F. Ursini, Antioxid. Redox Signaling, 2008, 10, 1485-1490. Peroxidase: A term of many meanings. 94. D. Koua, L. Cerutti, L. Falquet, C. J. A. Sigrist, G. Theiler, N. Hulo and C. Dunand, Nucleic Acids Res., 2009, 37, D261-D266. PeroxiBase: a database with new tools for peroxidase family classification.

105

95. B. L. Allen, P. D. Kichambare, P. Gou, I. I. Vlasova, A. A. Kapralov, N. Konduru, V. E. Kagan and A. Star, Nano Lett., 2008, 8, 3899-3903. Biodegradation of single-walled carbon nanotubes through enzymatic catalysis. 96. B. L. Allen, G. P. Kotchey, Y. Chen, N. V. K. Yanamala, J. Klein-Seetharaman, V. E. Kagan and A. Star, J. Am. Chem. Soc., 2009, 131, 17194-17205. Mechanistic investigations of horseradish peroxidase-catalyzed degradation of single-walled carbon nanotubes. 97. M. Salomon and F. Buchholz, Comparative biochemistry and physiology. Part B, Biochemistry & molecular biology, 2000, 125, 71-81. Effects of temperature on the respiration rates and the kinetics of citrate synthase in two species of Idotea (Isopoda, Crustacea). 98. M. Ghadermarzi and A. A. Moosavi-Movahedi, The Italian Journal of Biochemistry, 1997, 46, 197-205. The effects of temperature and pH on the kinetics of reactions between catalase and its suicide substrate hydrogen peroxide. 99. Y. Zhao, B. L. Allen and A. Star, J. Phys. Chem. A, 2011, 115, 9536-9544. Enzymatic degradation of multiwalled carbon nanotubes. 100. A. R. Sureshbabu, R. Kurapati, J. Russier, C. Menard-Moyon, I. Bartolini, M. Meneghetti, K. Kostarelos and A. Bianco, Biomaterials, 2015, 72, 20-28. Degradation-by-design: Surface modification with functional substrates that enhance the enzymatic degradation of carbon nanotubes. 101. G. P. Kotchey, B. L. Allen, H. Vedala, N. Yanamala, A. A. Kapralov, Y. Y. Tyurina, J. Klein-Seetharaman, V. E. Kagan and A. Star, ACS Nano, 2011, 5, 2098-2108. The enzymatic oxidation of graphene oxide. 102. Y. Li, L. Feng, X. Shi, X. Wang, Y. Yang, K. Yang, T. Liu, G. Yang and Z. Liu, Small, 2013, 2014, 10, 1544-1554. Surface coating-dependent cytotoxicity and degradation of graphene derivatives: towards the design of non-toxic, degradable nano-graphene. 103. V. R. Samygina, A. V. Sokolov, G. Bourenkov, M. V. Petoukhov, M. O. Pulina, E. T. Zakharova, V. B. Vasilyev, H. Bartunik and D. I. Svergun, PloS One, 2013, 8, e67145. Ceruloplasmin: macromolecular assemblies with iron-containing acute phase proteins. 104. S. J. Klebanoff, Proc. Assoc. Am. Physicians, 1999, 111, 383-389. Myeloperoxidase. 105. M. Zhang, M. Yang, C. Bussy, S. Iijima, K. Kostarelos and M. Yudasaka, Nanoscale, 2015, 7, 2834-2840. Biodegradation of carbon nanohorns in macrophage cells. 106. K. Bhattacharya, C. Sacchetti, R. El-Sayed, A. Fornara, G. P. Kotchey, J. A. Gaugler, A. Star, M. Bottini and B. Fadeel, Nanoscale, 2014, 6, 14686-14690. Enzymatic 'stripping' and degradation of PEGylated carbon nanotubes.

106

107. R. Gref, M. Lück, P. Quellec, M. Marchand, E. Dellacherie, S. Harnisch, T. Blunk and R. H. Müller, Colloids Surf., B, 2000, 18, 301-313. 'Stealth' corona-core nanoparticles surface modified by polyethylene glycol (PEG): Influences of the corona (PEG chain length and surface density) and of the core composition on phagocytic uptake and plasma protein adsorption. 108. Y. Ding, R. Tian, Z. Yang, J. Chen and N. Lu, Biophys. Chem., 2017, 222, 1-6. Effects of serum albumin on the degradation and cytotoxicity of single-walled carbon nanotubes. 109. Y. Ding, R. Tian, Z. Yang, J. Chen and N. Lu, Biophys. Chem., 2016, 218, 36-41. Binding of human IgG to single-walled carbon nanotubes accelerated myeloperoxidase-mediated degradation in activated neutrophils. 110. C. Le, C. Kunacheva and D. C. Stuckey, Environ. Sci. Technol., 2016, 50, 3074- 3081. “Protein” measurement in biological wastewater treatment systems: A critical evaluation. 111. M. Hadjidemetriou, Z. Al-Ahmady and K. Kostarelos, Nanoscale, 2016, 8, 6948-6957. Time-evolution of in vivo protein corona onto blood-circulating PEGylated liposomal doxorubicin (DOXIL) nanoparticles. 112. S. Sullivan, M. A. Salapow, R. Breen and D. H. Broide, Int. Arch. Allergy Immunol., 1994, 105, 150-154. Eosinophil peroxidase differs from neutrophil myeloperoxidase in Its ability to bind antineutrophil cytoplasmic antibodies reactive with myeloperoxidase. 113. F. T. Andon, A. A. Kapralov, N. Yanamala, W. Feng, A. Baygan, B. J. Chambers, K. Hultenby, F. Ye, M. S. Toprak, B. D. Brandner, A. Fornara, J. Klein-Seetharaman, G. P. Kotchey, A. Star, A. A. Shvedova, B. Fadeel and V. E. Kagan, Small, 2013, 9, 2721-2729, 2720. Biodegradation of single-walled carbon nanotubes by eosinophil peroxidase. 114. K. H. Steen, A. E. Steen and P. W. Reeh, J. Neurosci., 1995, 15, 3982-3989. A dominant role of acid pH in inflammatory excitation and sensitization of nociceptors in rat skin, in vitro. 115. J. A. Kellum, Critical Care, 2000, 4, 6-14. Determinants of blood pH in health and disease. 116. A. Singh and M. Agrawal, J. Environ. Biol. 2008, 29, 15-24. Acid rain and its ecological consequences. 117. D. Elgrabli, W. Dachraoui, C. Ménard-Moyon, X. J. Liu, D. Bégin, S. Bégin-Colin, A. Bianco, F. Gazeau and D. Alloyeau, ACS Nano, 2015, 9, 10113-10124. Carbon nanotube degradation in macrophages: live nanoscale monitoring and understanding of biological pathway.

107

118. Y. Ding, R. Tian, Z. Yang, J. Chen and N. Lu, J. Mater. Sci.: Mater. Med., 2017, 28, 7. NADPH oxidase-dependent degradation of single-walled carbon nanotubes in macrophages. 119. D. S. Arora, M. Chander and P. K. Gill, Int. Biodeterior. Biodegrad., 2002, 50, 115- 120. Involvement of lignin peroxidase, manganese peroxidase and laccase in degradation and selective ligninolysis of wheat straw. 120. N. Bansal and S. S. Kanwar, Sci. World J., 2013, 714639. Peroxidase(s) in environment protection. 121. G. Chandrasekaran, S.-K. Choi, Y.-C. Lee, G.-J. Kim and H. J. Shin, J. Ind. Eng. Chem., 2014, 20, 3367-3374. Biodegradation of single-walled carbon nanotubes in macrophages through respiratory burst modulation. 122. C. Zhang, W. Chen and P. J. J. Alvarez, Environ. Sci. Technol., 2014, 48, 7918- 7923. Manganese peroxidase degrades pristine but not surface-oxidized (carboxylated) single-walled carbon nanotubes. 123. G. Lalwani, W. Xing and B. Sitharaman, J. Mater. Chem. B, 2014, 2, 6354-6362. Enzymatic degradation of oxidized and reduced graphene nanoribbons by lignin peroxidase. 124. I. I. Vlasova, T. V. Vakhrusheva, A. V. Sokolov, V. A. Kostevich and A. A. Ragimov, J. Phys.: Conf. Ser., 2011, 291, 1-5. Peroxidase-induced degradation of single-walled carbon nanotubes: hypochlorite is a major oxidant capable of in vivo degradation of carbon. nanotubes. 125. S. Pan, N. P. Sardesai, H. Liu, D. Li and J. F. Rusling, Toxicol. Res. (Cambridge, U. K.), 2013, 2, 375-378. Assessing DNA damage from enzyme-oxidized single-walled carbon nanotubes. 126. D. C. Dale, L. Boxer and W. C. Liles, Blood, 2008, 112, 935-945. Assessing DNA Damage from Enzyme-Oxidized Single-Walled Carbon Nanotubes 127. C. Kantari, M. Pederzoli-Ribeil and V. Witko-Sarsat, Contrib. Microbiol., 2008, 15, 118-146. The role of neutrophils and monocytes in innate immunity. 128. C. Nathan, Nat. Rev. Immunol., 2006, 6, 173-182. Neutrophils and immunity: challenges and opportunities. 129. S. J. Klebanoff, J. Leukocyte Biol., 2005, 77, 598-625. Myeloperoxidase: friend and foe. 130. V. Brinkmann, U. Reichard, C. Goosmann, B. Fauler, Y. Uhlemann, D. S. Weiss, Y. Weinrauch and A. Zychlinsky, Science, 2004, 303, 1532-1535. Neutrophil extracellular traps kill bacteria. 131. G. Fanali, A. di Masi, V. Trezza, M. Marino, M. Fasano and P. Ascenzi, Mol. Aspects Med., 2012, 33, 209-290. Human serum albumin: from bench to bedside.

108

132. C. Farrera, K. Bhattacharya, B. Lazzaretto, F. T. Andon, K. Hultenby, G. P. Kotchey, A. Star and B. Fadeel, Nanoscale, 2014, 6, 6974-6983. Extracellular entrapment and degradation of single-walled carbon nanotubes. 133. M. Hadjidemetriou, Z. Al-Ahmady, M. Mazza, R. F. Collins, K. Dawson and K. Kostarelos, ACS Nano, 2015, 9, 8142-8156. In vivo biomolecule corona around blood-circulating, clinically used and antibody-targeted lipid bilayer nanoscale vesicles. 134. M. T. Silva and M. Correia-Neves, Front. Immunol., 2012, 3, 174. Neutrophils and macrophages: the main partners of phagocyte cell systems. 135. K. Inaba, M. Inaba, M. Deguchi, K. Hagi, R. Yasumizu, S. Ikehara, S. Muramatsu and R. M. Steinman, Proc. Natl. Acad. Sci. U. S. A., 1993, 90, 3038-3042. Granulocytes, macrophages, and dendritic cells arise from a common major histocompatibility complex class II-negative progenitor in mouse bone marrow. 136. S. Epelman, K. J. Lavine, A. E. Beaudin, D. K. Sojka, J. A. Carrero, B. Calderon, T. Brija, E. L. Gautier, S. Ivanov, A. T. Satpathy, J. D. Schilling, R. Schwendener, I. Sergin, B. Razani, E. C. Forsberg, W. Yokoyama, E. R. Unanue, M. Colonna, G. J. Randolph and D. L. Mann, Immunity, 2014, 40, 91-104. Embryonic and adult-derived resident cardiac macrophages are maintained through distinct mechanisms at steady state and during inflammation. 137. E. R. Unanue, Annu. Rev. Immunol., 1984, 2, 395-428. Antigen-presenting function of the macrophage. 138. V. E. Kagan, A. A. Kapralov, C. M. St. Croix, S. C. Watkins, E. R. Kisin, G. P. Kotchey, K. Balasubramanian, I. I. Vlasova, J. Yu, K. Kim, W. Seo, R. K. Mallampalli, A. Star and A. A. Shvedova, ACS Nano, 2014, 8, 5610-5621. Lung macrophages "digest" carbon nanotubes using a superoxide/peroxynitrite oxidative pathway. 139. M. Xu, J. Zhu, F. Wang, Y. Xiong, Y. Wu, Q. Wang, J. Weng, Z. Zhang, W. Chen and S. Liu, ACS Nano, 2016, 10, 3267-3281. Improved in vitro and in vivo biocompatibility of graphene oxide through surface modification: Poly(acrylic acid)-functionalization is superior to PEGylation. 140. D. Elgrabli, W. Dachraoui, H. d. Marmier, C. Ménard-Moyon, D. Bégin, S. Bégin- Colin, A. Bianco, D. Alloyeau and F. Gazeau, Sci. Rep., 2017, 7, 40997. Intracellular degradation of functionalized carbon nanotube/iron oxide hybrids is modulated by iron via Nrf2 pathway. 141. C. H. Wong, Z. Sofer, M. Kubesova, J. Kucera, S. Matejkova and M. Pumera, Proc. Natl. Acad. Sci. U. S. A., 2014, 111, 13774-13779. Synthetic routes contaminate graphene materials with a whole spectrum of unanticipated metallic elements.

109

142. M. Landry, M. Pinault, S. Tchankouo, E. Charon, A. Ridoux, J. Boczkowski, M. Mayne-L'Hermite and S. Lanone, Part. Fibre Toxicol., 2016, 13, 61. Early signs of multi-walled carbon nanotbues degradation in macrophages, via an intracellular pH- dependent biological mechanism; importance of length and functionalization. 143. D. Morgan, V. V. Cherny, R. Murphy, B. Z. Katz and T. E. DeCoursey, J. Physiol., 2005, 569, 419-431. The pH dependence of NADPH oxidase in human eosinophils. 144. T. I. Lam, A. M. Brennan-Minnella, S. J. Won, Y. Shen, C. Hefner, Y. Shi, D. Sun and R. A. Swanson, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, E4362-E4368. The pH dependence of NADPH oxidase in human eosinophils. 145. C. M. Girish, A. Sasidharan, G. S. Gowd, S. Nair and M. Koyakutty, Adv. Healthcare Mater., 2013, 2, 1489-1500. Confocal Raman imaging study showing macrophage mediated biodegradation of graphene in vivo. 146. F. Ginhoux, S. Lim, G. Hoeffel, D. Low and T. Huber, Front. Cell. Neurosci., 2013, 7, 45. Origin and differentiation of microglia. 147. A. Nimmerjahn, F. Kirchhoff and F. Helmchen, Science, 2005, 308, 1314-1318. Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. 148. E. S. Patchin, D. S. Anderson, R. M. Silva, D. L. Uyeminami, G. M. Scott, T. Guo, L. S. Van Winkle and K. E. Pinkerton, Environ. Health Perspect., 2016, 124, 1870-1875. Size-dependent deposition, translocation, and microglial activation of inhaled silver nanoparticles in the rodent nose and brain. 149. E. Ulvestad, K. Williams, S. Mork, J. Antel and H. Nyland, J. Neuropathol. Exp. Neurol., 1994, 53, 492-501. Phenotypic differences between human monocytes/macrophages and microglial cells studied in situ and in vitro. 150. D. L. Lefkowitz and S. S. Lefkowitz, Free Radicals Biol. Med., 2008, 45, 726-731. Microglia and myeloperoxidase: a deadly partnership in neurodegenerative disease. 151. R. N. Saha and K. Pahan, Antioxid. Redox Signaling, 2006, 8, 929-947. Regulation of inducible nitric oxide synthase gene in glial cells. 152. K. Dohi, H. Ohtaki, T. Nakamachi, S. Yofu, K. Satoh, K. Miyamoto, D. Song, S. Tsunawaki, S. Shioda and T. Aruga, J. Neuroinflammation, 2010, 7, 41. Gp91phox (NOX2) in classically activated microglia exacerbates traumatic brain injury. 153. A. E. Goode, D. A. Gonzalez Carter, M. Motskin, I. S. Pienaar, S. Chen, S. Hu, P. Ruenraroengsak, M. P. Ryan, M. S. P. Shaffer, D. T. Dexter and A. E. Porter, Biomaterials, 2015, 70, 57-70. High resolution and dynamic imaging of biopersistence and bioreactivity of etra and intracellular MWNTs exposed to microglial cells. 154. D. Mata, M. Amaral, A. J. S. Fernandes, B. Colaco, A. Gama, M. C. Paiva, P. S. Gomes, R. F. Silva and M. H. Fernandes, Nanoscale, 2015, 7, 9238-9251. Diels–

110

Alder functionalized carbon nanotubes for bone tissue engineering: in vitro/in vivo biocompatibility and biodegradability. 155. C. Bussy, C. Hadad, M. Prato, A. Bianco and K. Kostarelos, Nanoscale, 2016, 8, 590-601. Intracellular degradation of chemically functionalized carbon nanotubes using a long-term primary microglial culture model. 156. C. Guo, W. T. Al-Jamal, F. M. Toma, A. Bianco, M. Prato, K. T. Al-Jamal and K. Kostarelos, Bioconjugate Chem., 2015, 26, 1370-1379. Design of cationic multiwalled carbon nanotubes as efficient siRNA vectors for lung cancer xenograft eradication. 157. K. T. Al-Jamal, L. Gherardini, G. Bardi, A. Nunes, C. Guo, C. Bussy, M. A. Herrero, A. Bianco, M. Prato, K. Kostarelos and T. Pizzorusso, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 10952-10957. Functional motor recovery from brain ischemic insult by carbon nanotube-mediated siRNA silencing. 158. V. Neves, E. Heister, S. Costa, C. Tîlmaciu, E. Borowiak-Palen, C. E. Giusca, E. Flahaut, B. Soula, H. M. Coley, J. McFadden and S. R. P. Silva, Adv. Funct. Mater., 2010, 20, 3272-3279. Cellular localization, accumulation and trafficking of double- walled carbon nanotubes in human prostate cancer cells. 159. J. Hou, B. Wan, Y. Yang, X. M. Ren, L. H. Guo and J. F. Liu, Int. J. Mol. Sci., 2016, 17, 409. Biodegradation of single-walled carbon nanotubes in macrophages through respiratory burst modulation. 160. C. A. Umscheid, D. J. Margolis and C. E. Grossman, Postgrad. Med., 2011, 123, 194-204. Key concepts of clinical trials: A narrative review. 161. D. Elgrabli, M. Floriani, S. Abella-Gallart, L. Meunier, C. Gamez, P. Delalain, F. Rogerieux, J. Boczkowski and G. Lacroix, Part. Fibre Toxicol., 2008, 5, 20. Biodistribution and clearance of instilled carbon nanotubes in rat lung. 162. A. A. Shvedova, A. A. Kapralov, W. H. Feng, E. R. Kisin, A. R. Murray, R. R. Mercer, C. M. St Croix, M. A. Lang, S. C. Watkins, N. V. Konduru, B. L. Allen, J. Conroy, G. P. Kotchey, B. M. Mohamed, A. D. Meade, Y. Volkov, A. Star, B. Fadeel and V. E. Kagan, PloS One, 2012, 7, e30923. Impaired clearance and enhanced pulmonary inflammatory/fibrotic response to carbon nanotubes in myeloperoxidase-deficient mice. 163. N. Shinohara, T. Nakazato, K. Ohkawa, M. Tamura, N. Kobayashi, Y. Morimoto, T. Oyabu, T. Myojo, M. Shimada, K. Yamamoto, H. Tao, M. Ema, M. Naya and J. Nakanishi, J. Appl. Toxicol., 2016, 36, 501-509. Long-term retention of pristine multiwalled carbon nanotubes in rat lungs after intratracheal instillation. 164. A. Nunes, C. Bussy, L. Gherardini, M. Meneghetti, M. A. Herrero, A. Bianco, M. Prato, T. Pizzorusso, K. T. Al-Jamal and K. Kostarelos, Nanomedicine (Lond), 2012,

111

7, 1485-1494. In vivo degradation of functionalized carbon nanotubes after stereotactic administration in the brain cortex. 165. Y. Sato, A. Yokoyama, Y. Nodasaka, T. Kohgo, K. Motomiya, H. Matsumoto, E. Nakazawa, T. Numata, M. Zhang, M. Yudasaka, H. Hara, R. Araki, O. Tsukamoto, H. Saito, T. Kamino, F. Watari and K. Tohji, Sci. Rep., 2013, 3, 2516. Long-term biopersistence of tangled oxidized carbon nanotubes inside and outside macrophages in rat subcutaneous tissue. 166. A. Sasidharan, S. Swaroop, C. K. Koduri, C. M. Girish, P. Chandran, L. S. Panchakarla, V. H. Somasundaram, G. S. Gowd, S. Nair and M. Koyakutty, Carbon, 2015, 95, 511-524. Comparative in vivo toxicity, organ biodistribution and immune response of pristine, carboxylated and PEGylated few-layer graphene sheets in Swiss albino mice: A three month study. 167. D. Elgrabli, M. Floriani, S. Abella-Gallart, L. Meunier, C. Gamez, P. Delalain, F. Rogerieux, J. Boczkowski and G. Lacroix, Part. Fibre Toxicol., 2008, 5, 1743-8977. 168. G. Mastrangelo, E. Fadda and V. Marzia, Environ. Health Perspect., 1996, 104, 1166-1170. Biodistribution and clearance of instilled carbon nanotubes in rat lung. 169. Y. V. Pashin and L. M. Bakhitova, Environ. Health Perspect., 1979, 30, 185-189. Mutagenic and carcinogenic properties of polycyclic aromatic hydrocarbons. 170. M. Kendall and I. Lynch, Nat. Nanotechnol., 2016, 11, 206-210. Long-term monitoring for nanomedicine implants and drugs. 171. T. Lammers, F. Kiessling, W. E. Hennink and G. Storm, J. Controlled Release, 2012, 161, 175-187. Drug targeting to tumors: principles, pitfalls and (pre-) clinical progress. 172. Z. Liu, K. Chen, C. Davis, S. Sherlock, Q. Cao, X. Chen and H. Dai, Cancer Res., 2008, 68, 6652-6660. Drug delivery with carbon nanotubes for in vivo cancer treatment. 173. J. Tian, Y. Luo, L. Huang, Y. Feng, H. Ju and B.-Y. Yu, Biosens. Bioelectron., 2016, 80, 519-524. Pegylated folate and peptide-decorated graphene oxide nanovehicle for in vivo targeted delivery of anticancer drugs and therapeutic self-monitoring. 174. K. Fizazi and L. Zelek, Annals of oncology : Official Journal of the European Society for Medical Oncology, 2000, 11, 133-149. Is òne cycle every three or four weeks' obsolete? A critical review of dose-dense chemotherapy in solid neoplasms. 175. F. Ordikhani, M. Erdem Arslan, R. Marcelo, I. Sahin, P. Grigsby, J. K. Schwarz and A. K. Azab, Pharmaceutics, 2016, 8, 23. Drug Delivery Approaches for the Treatment of Cervical Cancer.

112

176. E. Byun and H. Lee, J. Nanosci. Nanotechnol., 2014, 14, 7395-7401. Enhanced loading efficiency and sustained release of doxorubicin from hyaluronic acid/graphene oxide composite hydrogels by a mussel-inspired catecholamine. 177. J. Hong, N. J. Shah, A. C. Drake, P. C. DeMuth, J. B. Lee, J. Chen and P. T. Hammond, ACS Nano, 2012, 6, 81-88. Graphene multilayers as gates for multi-week sequential release of proteins from surfaces. 178. A. Bianco, K. Kostarelos and M. Prato, Chem. Commun., 2011, 47, 10182-10188. Making carbon nanotubes biocompatible and biodegradable. 179. M. Saffarzadeh and K. T. Preissner, Curr. Opin. Hematol., 2013, 20, 3-9. Fighting against the dark side of neutrophil extracellular traps in disease: manoeuvres for host protection. 180. D. L. Laskin, V. R. Sunil, C. R. Gardner and J. D. Laskin, Annu. Rev. Pharmacol. Toxicol. 2011, 51, 267-288. Macrophages and tissue injury: agents of defense or destruction? 181. J. Zhang, H. Yang, G. Shen, P. Cheng, J. Zhang and S. Guo, Chem. Commun., 2010, 46, 1112-1114. Reduction of graphene oxide via L-ascorbic acid. 182. R. Langer and J. P. Vacanti, Science 1993, 260, 920-926. Tissue engineering. 183. F. J. O'Brien, Mater.Today, 2011, 14, 88-95. Biomaterials and scaffolds for tissue engineering. 184. D. W. Hutmacher, Biomaterials, 2000, 21, 2529-2543. Scaffolds in tissue engineering bone and cartilage. 185. F. T. Moutos and F. Guilak, Biorheology, 2008, 45, 501-512. Composite scaffolds for cartilage tissue engineering. 186. S. R. Shin, Y.-C. Li, H. L. Jang, P. Khoshakhlagh, M. Akbari, A. Nasajpour, Y. S. Zhang, A. Tamayol and A. Khademhosseini, Adv. Drug Delivery Rev. 2016, 105, Part B, 255-274. Graphene-based materials for tissue engineering. 187. C. D. Hoemann, J. Sun, A. Légaré, M. D. McKee and M. D. Buschmann, Osteoarthritis and Cartilage, 2005, 13, 318-329. Tissue engineering of cartilage using an injectable and adhesive chitosan-based cell-delivery vehicle. 188. L. Yildirimer, A. Buanz, S. Gaisford, E. L. Malins, C. Remzi Becer, N. Moiemen, G. M. Reynolds and A. M. Seifalian, Sci. Rep., 2015, 5, 15040. Controllable degradation kinetics of POSS nanoparticle-integrated poly(ε-caprolactone urea)urethane elastomers for tissue engineering applications. 189. O. V. Horst, M. G. Chavez, A. H. Jheon, T. Desai and O. D. Klein, Dent. Clin. North Am. 2012, 56, 495-520. Stem cell and biomaterials research in dental tissue engineering and regeneration.

113

190. S. Ravi and E. L. Chaikof, Regener. Med., 2010, 5, 107. Biomaterials for vascular tissue engineering. 191. J. Zhu and R. E. Marchant, Expert Rev. Med. Devices, 2011, 8, 607-626. Design properties of hydrogel tissue-engineering scaffolds. 192. R. G. Ellis-Behnke, Y. X. Liang, S. W. You, D. K. Tay, S. Zhang, K. F. So and G. E. Schneider, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 5054-5059. Nano neuro knitting: Peptide nanofiber scaffold for brain repair and axon regeneration with functional return of vision. 193. T. S. Sreeprasad and V. Berry, Small, 2013, 9, 341-350. How do the electrical properties of graphene change with its functionalization? 194. C. G. Núñez, W. T. Navaraj, E. O. Polat and R. Dahiya, Adv. Funct. Mater., 2017, 27, 1606287. Energy-Autonomous, Flexible, and Transparent Tactile Skin. 195. K. Chen and X. Chen, Current topics in medicinal chemistry, 2010, 10, 1227-1236. Design and development of molecular imaging probes 196. A. Dellinger, J. Olson, K. Link, S. Vance, M. G. Sandros, J. Yang, Z. Zhou and C. L. Kepley, Journal of cardiovascular magnetic resonance, 2013, 15, 7. Functionalization of gadolinium metallofullerenes for detecting atherosclerotic plaque lesions by cardiovascular magnetic resonance. 197. M. Rogosnitzky and S. Branch, Biometals, 2016, 29, 365-376. Gadolinium-based contrast agent toxicity: a review of known and proposed mechanisms. 198. Z. Yue, P. Lv, H. Yue, Y. Gao, D. Ma, W. Wei and G. Ma, Chem. Commun., 2013, 49, 3902-3904. Inducible graphene oxide probe for high-specific tumor diagnosis. 199. T. C. Liu, M. C. Chuang, C. Y. Chu, W. C. Huang, H. Y. Lai, C. T. Wang, W. L. Chu, S. Y. Chen and Y. Y. Chen, ACS Appl. Mater. Interfaces, 2016, 8, 187-196. Implantable graphene-based neural electrode interfaces for electrophysiology and neurochemistry in in vivo hyperacute stroke model.

114

Chapter 3

3. Hypothesis, Aims and Objectives

The experimental work contained within this thesis has been specifically designed to interrogate the biological interactions and fate of the oxidised derivative of graphene, graphene oxide (GO). As mentioned, this material has been known to exist for decades. Following the successful isolation of graphene, its pristine counterpart, interest has re- emerged regarding this unique material. This is exemplified by the increasingly populous literature studies exploring the usage of this material for biomedical and non-biomedical applications alike. Fittingly, a detailed and systematic set of studies is described to address one of the key aspects of the materials biological fate following its in vivo exposure to living subjects, which is it’s in vivo biodegradability. In the literature this has not been directly addressed.

The overarching hypothesis of this work is: Following administration in vivo, GO sheets will be subject to biodegradative processes which will result in the efficient destruction of the material. This degradation will be mediated predominantly by macrophages. Macrophages are long lived and key cells, that in vivo are widely implicated in the removal of endogenous debris as well as exogenous materials, which includes nano-particulates including sp2 hybridised carbon nanomaterials (CNMs).

In order to test this hypothesis, a series of experiments with the following objectives was designed :

 To characterise and define the degradative changes that GO undergoes over time, following its exposure to a highly oxidising chemical (1%NaClO) in comparison to other CNMs. This will better inform us about the detectable changes that the CNMs undergo that are characteristic of degradation (Chapter 5).  To track the in vivo biodegradability of GO nanosheets in the mice spleen following the intravenous administration of the material at an optimised dosage. The Nanomedicine lab’s previous work has identified the spleen to be the principal organ involved in the accumulation of the material over the short term. In addition to this main objective, the toxicological implications of this accumulation and the specific cell sub populations that are responsible, if any, for this accumulation as well as any biodegradative processing were identified (Chapter 6).

115

 To characterise the dynamics and the extent that thoroughly characterised GO sheets of different lateral dimensions can translocate from the nose to the brain following their administration via intranasal instillation. In addition, it was aimed to identify the cell populations, if any, that are responsible for accumulating the translocated material and to interrogate the biodegradability of these materials over time within the brain (Chapter 7).

116

Chapter 4

4. Hypochlorite Degrades 2D Graphene Oxide Sheets Faster than 1D Oxidised Carbon Nanotubes and Nanohorns

Leon Newman1, Neus Lozano1, Minfang Zhang2, Sumio Iijima2, Masako Yudasaka2, Cyrill Bussy1,* and Kostas Kostarelos1,*

1 Nanomedicine Laboratory, Faculty of Biology, Medicine and Health and National Graphene Institute, University of Manchester, AV Hill Building, Manchester M13 9PT, United Kingdom

2 Institute of Advanced Science and Industrial Technology (AIST) Tsukuba 305-8565, Ibaraki, Japan

4.1. Statement. Leon Newman, Dr Cyrill Bussy and Prof Kostas Kostarelos planned the study, and designed the experiments. Leon Newman wrote the full manuscript and performed all the analytical experiment except for the XPS, which was performed at the NEXUS facility at the University of Newcastle and the generated data analysed by Dr Neus Lozano. Dr Neus Lozano also synthesised the graphene oxide, Dr Jeroen Van den Bossche synthesised the oxidised carbon nanotubes. Minfang Zhang, Sumio Iijima and Masako Yudasaka synthesised the oxidised carbon nanohorns. This manuscript has been critically reviewed by all authors and has been NPJ 2D Materials and Applications, doi:10.1038/s41699-017-0041-3.

______

* Correspondence to either: [email protected]

[email protected]

117

4.2. Abstract.

Carbon nanostructures are currently fuelling a revolution in science and technology in areas ranging from aerospace engineering to electronics. Oxidised carbon nanomaterials, such as graphene oxide, exhibit dramatically improved water dispersibility compared to their pristine equivalents, allowing their exploration in biology and medicine. Concomitant with these potential healthcare applications, the issue of degradability has been raised and has started to be investigated. The aim of the present study was to assess the potential of hypochlorite, a naturally occurring and industrially used ion, to degrade oxidised carbon nanomaterials within a week. Our main focus was to characterise the physical and chemical changes that occur during the degradation of graphene oxide compared to two other oxidised carbon nanomaterials, namely carbon nanotubes and carbon nanohorns. The kinetics of degradation were closely monitored over a week using a battery of techniques including visual observation, UV-vis spectroscopy, Raman spectroscopy, infrared spectroscopy, transmission electron microscopy and atomic force microscopy. Graphene oxide was rapidly degraded into a dominantly amorphous structure lacking the characteristic Raman signature and microscopic morphology. Oxidised carbon nanotubes underwent degradation via a wall exfoliation mechanism, yet maintained a large fraction of the sp2 carbon backbone, while the degradation of oxidised carbon nanohorns was somewhat intermediate. The present study shows the timeline of physical and chemical alterations of oxidised carbon nanomaterials, demonstrating a faster degradation of 2D graphene oxide sheets compared to 1D oxidised carbon nanomaterials over 7 days in the presence of an oxidising species.

118

4.3. Introduction.

Graphene is a two dimensional sp2 carbon-based material that has acquired the attention and focus of experts from a myriad of disciplines. This comes following the revelation of its exceptional electronic1, mechanical2, thermal3 and optical properties4. Among the various derivatives of graphene, graphene oxide (GO)5 has demonstrated tremendous utility in water-based environments, since it surmounts one of graphene’s caveats for its application in biomedicine, which is hydrophobicity. As of late, there are increasing proof-of-concept studies exploring the applicability of GO for biomedical use in comparison to other pristine and derivatised sp2 hybridised carbon nanomaterials (CNMs), the body being predominantly a hydrophilic environment. For example, the delivery of both drug and diagnostic molecules6, the capturing of circulating cancer cells7, and the photothermal ablation of cancers8. Other studies have proposed the use GO for its antibacterial properties9 or as scaffolds in tissue engineering10. Looking less at the biomedical applicability and more at everyday applications, studies have demonstrated GOs usage, in its reduced form, in paints as a means to prevent metal corrosion11 or as molecular sieves in water filtration systems for desalination12. Overall these studies demonstrate the potential for wide use of GO in many commercial products.

One of the necessities for the introduction of a new material for industrial, medical or lay use is to understand its fate over time and therefore its potential degradability, for example in the event of an intentional or unintentional release into the environment or human exposure. In relation to the toxicological consequences of CNMs, such as graphene, on living organisms there are still inconclusive results13. Some studies suggest the ability of graphene, for example, to penetrate cell membranes and potentially damage the cell interior14, while others report biocompatibility15. Further studies are therefore warranted in this area to conclude on the toxicological profile of CNMs. Regarding degradability, several studies have investigated and demonstrated the degradation of CNMs including graphene using enzymes such as lignin peroxidase16 or horseradish peroxidase17. Chemical strategies such as Fenton-based chemistry have also been tested18. However, studies directly comparing the degradation kinetics of CNMs of different geometries in an oxidative environment are lacking.

In this study, it was the aim to further the current understanding of the fate of oxidised CNMs in an oxidative environment. We tested whether sodium hypochlorite (NaClO), colloquially known as bleach, was able to efficiently degrade GO flakes in suspension. Comparing the

119 chemical degradation of GO to that of two other oxidised CNMs, specifically carbon nanohorns (OxNH) and oxidised multiwall carbon nanotubes (OxMWNT), we reasoned that GO will degrade faster than the other two nanomaterials due to its unique physicochemical features. NaClO was chosen since it is a commonly used chemical by the public and industry. Moreover, hypochlorite (ClO-) is naturally produced in the human body by various enzymes such as myeloperoxidase and eosinophil peroxidase19. The hypothesis was that the strong oxidative action of ClO-, from NaClO, would induce oxidative damage to the graphitic backbone, ultimately degrading the nanostructures to various extents. The degradation processes mediated by ClO- were followed over a week using a battery of characterisation methods. These included visual observation, transmission electron microscopy (TEM), atomic force microscopy (AFM), Raman spectroscopy, UV-vis spectroscopy and Fourier transform infrared spectroscopy (FT-IR). The observations were compared against when the materials were incubated in water. It was found that incubation in NaClO induced severe structural modifications in GO that was consistent with the materials degradation. Over the course of a week, GO degraded more rapidly than OxNH or OxMWNT. The results of this study adds information to ratify a proposal we made in a previous report concerning the mechanism by which carbon-based nanomaterials may degrade under a strong oxidative environment20, 13.

120

4.4. Results.

Figure 4.1 shows the characterisation of the as synthesised GO, OxMWNT and OxNH. The Raman spectra obtained demonstrate an enhanced scattering intensity in the D band relative to the G band, indicative of the presence of defects, expected as the materials were oxidised. AFM images show an average height of 2.3 ± 1.1 nm, 35.2 ± 5.0 nm and 65 ± 10 nm for GO, OxMWNT and OxNH structures, respectively. The average lateral dimensions (taken in the longest dimensions were determined from AFM analyses) of GO, OxMWNT and OxNH were shown to be 900 ± 500 nm, 800 ± 200 nm and 125 ± 35 nm. TEM imaging evidenced the characteristic shapes of the nanomaterials and confirms the dimensions indicated by AFM. The XPS and TGA (Figure 4.S2) in a combinatory fashion confirmed that all three nanomaterials contained oxygen functionalities.

121

GO OxMWNT OxNH

ID/IG= 1.36 + 0.01 I /I = 1.75 + 0.03 ID/IG = 1.95 + 0.08 n=5 D G

n = 5 n=5

Raman

RamanIntensity a.u. /

Raman Intensity / a.u. / Intensity Raman Raman Intensity / / Intensity Raman a.u. 1000 2000 3000 1000 2000 3000 1000 2000 3000 -1 Raman Shift / cm-1 Raman Shift / cm Raman Shift / cm-1

10nm 40 nm 80 nm

0 nm 0 nm 0 nm

AFM TEM

Total System π- π* O-C=O C=O C-O-C C-OH C-C&C=C Functionalisation (%) 289.9 eV - - - 285.6 eV 284.6 eV Graphite 15.7% - - - 19.2% 65.1% 19.2% 290.8 eV 288.0 eV 286.8 eV 286.2 eV - 284.6 eV GO 1.2% 16.2% 26.3% 19.2% - 37.2% 61.7% 289.7 eV - - - 285.7 eV 284.6 eV MWNT 17.2% - - - 20.5% 62.3% 20.5%

XPS - - 287.2 eV - 285.3 eV 284.6 eV OxMWNT - - 19.8% - 22.4% 57.8% 42.2% 289.3 eV - - - 285.6 eV 284.6 eV NH 21.2% - - - 27.1% 51.7% 27.1% - 288.2 eV - - 285.6 eV 284.6 eV OxNH - 17.8% - - 24.4% 57.8% 42.2%

Figure 4.1 Characterisation of the starting oxidised carbon nanomaterials.

Carbon nanomaterials: graphene oxide (GO), oxidised multiwalled nanotubes (OxMWNT) and oxidised nanohorns (OxNH) were characterised by Raman spectroscopy to demonstrate the defected nature of the starting materials. AFM (scale bars = 500 nm) and TEM (scale bars = 100 nm) were utilised to allow visualisation of the nanomaterials and to indicate the dimensions of the materials. XPS was used to indicate the relative degree of oxygen functionalities on each of the nanostructures. XPS survey spectra and TGA are shown in the Supplementary Information (Figures 4.S1 and 4.S2, respectively) (XPS was conducted at the NEXUS facility and analysed by Dr Neus Lozano).

122

Figure 4.2 shows optical images (Figure 4.2A) and UV-vis absorbance at 700 nm (Figure

4.2B). Images show how the colour of the dispersions of GO, OxMWNT and OxNH in H2O and 1 % NaClO vary over 7 days. Only the dispersions containing carbon nanomaterials in 1 % NaClO demonstrated a change in colour. This was true for all of the CNMs but more so for GO and OxMWNT, with OxNH only showing a modest increase of transparency by Day

7. For CNMs that were dispersed in H2O, the contrary was true, no decrease in colour was apparent. UV-vis spectroscopy, demonstrated a decrease in absorbance at 700 nm over time for OxMWNT and OxNH when dispersed in 1 % NaClO (Figure 4.2B). The absorption signal decreased at a greater rate for OxMWNT than for OxNH and did not proceed to reduce further than the value obtained at Day 5. The signal for GO did not show absorption in this region. Moreover, it was not possible to infer information from the characteristic absorption peak at 230 nm due to interference by the NaClO components21. The absorption of all CNMs placed in H2O, did not differ during the course of the experiment, indicating their maintenance. Spectra0.4 showing absorption between 400 – 800 nm are given in Figure 4.S3.

A 0 B

0.3 Day Day

1 0.4 H2O

GO H2O Day Day 0.2 Ox-MWNT H2O 0.3 Ox-NH H2O NaClO 1% Day 2 Day GO NaClO 1% 0.2 3 Ox-MWNT NaClO 1%

0.1 Ox-NH NaClO 1% Day Day

0.1

Absorbance700 at nm Absorbance700nm at Day 5 Day 0.0 0.0 0 1 20 13 2 4 3 45 5 6 6 77

Day 7 Day Time / DaysTime / Days In H2O In NaClO Controls Figure 4.2 Optical changes of oxidised carbon nanomaterials over time in 1% NaClO.

(A) Visual appearance and colour of carbon nanomaterials in H2O and in 1% NaClO at time points

Day 0, 1, 2, 3, 5 and 7 are shown with controls: 1 % NaClO and H2O alone. The colour of oxidised carbon nanomaterials became less obvious over time when dispersed in 1 % NaClO (samples were all repeated in triplicate). (B) UV-vis was measured at 700 nm for all samples to indicate the degree of light absorption and scattering indicative of the physicochemical state of the dispersions (standard error is presented for each point, where n = 3). For OxMWNT and OxNH, there is a decrease in intensity at 700 nm.

Raman spectroscopy was employed to yield quantifiable data regarding the structural perturbations in the CNMs when incubated in 1 %NaClO. Figure 4.3 shows the evolution of the Raman spectra of GO over time in H2O (Figure 4.3A) and 1 % NaClO (Figure 4.3B). At Day 0 in 1 %NaClO, it is clear that characteristic scattering in the D and G bands are present

123 in the Raman spectra of all investigated CNMs. Raman scattering peaks were observed in the region of the D and G bands present at ~1330 cm-1 and ~1590 cm-1 respectively for all three nanomaterials indicating their graphitic backbone. The positions of these scattering peaks remained constant, however their relative intensities altered over time. The D band scattering intensity appears to increase relative to that of the G band for all materials although to different extents as seen in Figure 4.3C. By Day 3 and Day 5 for GO and OxNH respectively, the ratio of the scattering intensity in the D relative to the G bands (ID/IG) decreased (Figure 4.3C). For OxMWNT however, the characteristic features remained intact, although with a progressively increasing ID/IG ratio. The corresponding spectra of carbon nanomaterials dispersed in H2O are shown in Figure 4.S6-S8.

124

A GO OxMWNT OxNH

B GO OxMWNT OxNH

C 2.5

G 2.0

/I D I GO OxMWNT OxNH 1.5

2.5 1.0 0 1 2 32.0 4 5 6 7 Time / Days 1.5

1.0 Figure 4.3 Evolution of Raman spectra of oxidised 0carbon1 2 nanomaterials.3 4 5 6 7

The progressive changes in the spectral features of GO, OxMWNT and OxNH are shown at Days 0,

1, 2, 3, 5 and 7, when incubated in (A) H2O and (B) 1 % NaClO. At each time point the spectra is given in a different colour for clarity. (C) The average ID/IG ratio over time for each CNM (GO-Green, OxMWNT-Blue and OxNH- Red) when incubated in 1 % NaClO is given, where discernible, in C. Error bars have also been presented for each material in the respective colours at each time point (n = 30 measurements). All averaged individual Raman spectra have been presented in Figure 4.S6- 5.S8.

125

Figure 4.4 shows the ultra-structural changes in the CNMs over time when incubated in 1 % NaClO, as followed by TEM. It is evident from the representative images of GO, OxMWNT and OxNH that each nanomaterial underwent gross structural and morphological changes, that were progressive over the course of the 7 days incubation in 1 %NaClO. Images of the corresponding CNMs when dispersed in H2O are displayed in Figure 4.S4. Atomic force microscopy (AFM) was used to provide corroborative evidence of the changes in the different CNMs structure and morphology when incubated in 1 % NaClO (Figure 4.4). They evidenced similar changes to that shown via TEM. An imaging area of 4 µm2 was chosen in order to provide evidence of the ultra-structural changes that occurred over the course of the 7 days. TEM AFM

GO OxMWNT OxNH GO OxMWNT OxNH

Day 0 Day

Day 1 Day

Day 2 Day

Day 3 Day

Day 5 Day Day 7 Day

Figure 4.4 Representative observations of ultra-structural changes in oxidised carbon nanomaterials exposed to 1 % NaClO as detected by TEM (left) and AFM (right) over time.

Samples of GO, OxMWNT and OxNH were separated from NaClO salts via centrifugation at Days 0, 1, 2, 3, 5 and 7, and analysed via TEM (100 kV, scale bars = 100 nm) and AFM in tapping mode (lateral scale bars = 400 nm, height scale bars = 0 - 5 nm for GO, 0 - 20 nm for both OxMWNT and OxNH).

126

4.5. Discussion.

NaClO when placed in water results in the generation of ClO- and sodium hydroxide (NaOH). ClO- is a powerful oxidising agent. During an oxidation reaction, the ClO- donates an oxygen atom and its electrons to the species to be oxidised. This releases the chloride ion (Cl-). It is this process that allows NaClO to be used as an antimicrobial22, bleaching or strong oxidising agents. NaClO can also be used in chemical synthesis23. Nature too uses hypochlorous acid (HClO) and ClO- for example, in immune defence where it is produced by activated neutrophils and macrophages24. Within these cells, enzymes such as

- myeloperoxidase catalyse the conversion of H2O2 to HClO / ClO . These products are able to kill invading pathogens directly as well as modifying specific extracellular features25,26 which can further promote the immune response allowing the effective removal of invading pathogens.

With the recent surge of interest in the use of oxidised CNMs and the demonstration that these materials can undergo HClO / ClO- (from peroxidase enzymes) mediated biodegradation19, it was our intention to use NaClO in water, hence ClO-, in order to follow the oxidative degradation of different CNMs over a week. Previous studies have demonstrated the effect of ClO- on carbon nanotubes27-30. The authors utilised the action of

- myeloperoxidase to convert H2O2 to ClO , which subsequently acted to oxidise the carbon nanotubes and ultimately degrade them. The oxidative action of ClO- was confirmed, as aqueous NaClO was used as a positive control to degrade the CNMs. However, a direct comparison of the oxidative degradation kinetics of different CNMs is lacking. The specific objectives of the present study were therefore to assess the capacity of ClO- to induce structural deformations on GO compared to two other oxidised CNMs, and to compare the structural changes and kinetics of decay. A concentration of 1 % NaClO was selected as it is comparable to that used for house hold applications (3 - 6 %) and also could degrade the materials at a rate that allowed us to follow the gradual structural changes that occurred in the nanomaterials over a week, as we determined through pilot studies.

To achieve our aims, a battery of characterisation techniques was employed. Characterisation of the starting oxidised CNMs revealed that all three nanomaterials were initially functionalised to a similar extent as indicated by TGA albeit with some differences with GO being the most and OxMWNT being the least functionalised (Figure 4.S2). This trend was corroborated by the XPS data. It appeared however, that the XPS results indicated that there was a slightly larger difference between the functionalisation of GO

127 relative to the other two nanomaterials (Figure 4.S1) than suggested by TGA. This may be because XPS is a surface technique that only allows a detailed regional surface analysis (analysis depth ≈ 10 nm)31, as opposed to TGA which considers the entire sample structure allowing a more global analysis32. This is especially important when analysing OxMWNT and OxNH dahlia like structures, which have internal and external structural components, compared to 2D GO which does not. Despite these differences however, the results suggest that the overall functionalisation of all three materials was comparable. Importantly, though the total functionalisation was comparable, the ratio of the specific surface oxygen functionalities differed between the respective nanostructures. GO contained a mixture of carboxylic acid, carbonyl and epoxide groups, whereas the OxMWNT contained carbonyl and hydroxyl groups, whilst OxNH was shown to contain carboxylic acid and hydroxyl groups. This may be a consequence of the starting nanomaterials dimensionality and also the means by which they were oxidised. It should be considered however, that the resolution of the C1s spectra was too low to allow for the deconvolution of hydroxyls from epoxides in GO. Nevertheless the GO is likely to also contain hydroxyl groups as per its FT-IR spectra (Figure 4.S9).

Other characterisation techniques used such as Raman spectroscopy (Figure 4.1) demonstrated the presence of defects in the crystalline network of the starting materials as shown by an elevated D band scattering, which is expected following an oxidative reaction. The Raman spectral fingerprints obtained for the synthesised GO, OxMWNT and OxNH were comparable to that reported previously in the literature using identical or similar oxidation processes15,33.

To then monitor degradation, other characterisation techniques were used including visual observation (Figure 4.2A) and UV-vis absorption (Figure 4.2B) to gain superficial information relating to the amount of material present in solution over time when the materials were incubated in 1% NaClO. It was interesting to observe that the GO dispersions were of a yellow/brown colour whilst the OxMWNT and OxNH had a black colouration. This has also been demonstrated by previous reports in the literature15,33. An explanation for the colouration is that during synthesis, the oxidation of the pristine graphite removes carbon atoms from participating in the extensive electron delocalisation present in graphene. This initially opens the band gap of the component graphene in the longer wavelength regions of the visible spectrum as a result of the reduction in electron delocalisation. Depending on the extent of oxidation, the band gap will continue to correspond to shorter wavelengths of light as the paths for electron delocalisation become increasingly limited due to further disruption

128 of the π system. The lowest unoccupied C=C π electron states increase in energy as their abundance decreases, resulting in the change in colour from black graphene / graphite to orange-yellow GO dispersions34. The yellow brown colour cannot be seen obviously for OxMWNT and OxNH because of the denser structures, as well as lower degrees of oxidation in enclosed features such as the intact internal walls of OxMWNT, which may be oxidised to lesser extents. Following incubation in 1 % NaClO, the intensities of the colours for each dispersion reduced in comparison to when the CNMs were incubated in water. This suggests a decrease in the presence of intact CNMs. On ceasing of magnetic stirring to image the contents in the reaction vessels, we noticed that the degrading OxMWNT gradually came out of solution to some extent over time, which never occurred prior to Day 3. This is potentially indicative of the decay of the outer walls and the revealing of more pristine and hydrophobic features of the inner walls. All other measurements were therefore completed following re-dispersion via pipetting to avoid inconsistent measurements.

UV-vis spectra measurements were performed at 700 nm for all tested CNMs (Figure 4.2B). We did not use the absorption maxima of the tested carbon nanomaterials due to the interfering light absorption by the oxidative ClO-21. Instead, we selected the 700 nm wavelength in order to measure simultaneously the light scattering and absorption by each material, as a means to spectroscopically evaluate the amount of materials present. For both OxMWNT and OxNH, there was a clear decrease of absorption and scattering overtime. This suggests a progressive but continuous degradation of both nanomaterials by ClO-. The decrease in absorption and scattering by OxNH could be explained by a slow destruction of the starting dahlia-like shape as the individual nanohorns become progressively more oxidised. The decrease in scattering by OxMWNT was less regular, suggesting more complex degradation kinetics as previously discussed19,20. In contrast, no changes in absorption or scattering characteristics were observable for GO. Although useful for studying the patterns of CNM degradation, UV-vis is not ideal for studying GO, especially in our case, where components of the ClO- absorb near the GO absorption maximum21 and no absorption or scattering is seen at longer wavelengths.

In order to obtain more conclusive quantifiable evidence of degradation, Raman spectroscopy was employed (Figure 4.3A-C). All spectral intensities were normalised to the -1 primary in plane vibrational mode peak intensity at 1590 cm (G band). The value of the ID/IG ratio was then calculated. The D peak, present at ~1330 cm-1, provided information about the abundance of defects35. As defects increase in the graphene plane, the D band scattering intensity will increase relative to the G band scattering intensity (ID/IG) up to a maximum

129 defect density or minimum crystal domain size in accordance with the Tuinstra-Koenin (TK) equation. Therefore until this maximum defect density the degree of disorder can be rationalised36,37,38.

Over the first day, all the carbon nanomaterials underwent a slight increase in ID/IG (Figure 4.3C), indicative of the increasing degree of defects in the graphene planes due to oxidation by ClO-. Similar observations were demonstrated in other carbon nanomaterial 17 degradation studies . It is interesting in this early stage (Days 0 - 2) the ID/IG ratio only marginally increased. This could due to the already highly defected nature of the exposed layers of the materials and perhaps the presence of oxidised debris derived from the synthesis that was physisorbed to the surface of the sheet. These structures may have provided some initial protection to the underlying material from the oxidising ClO-. The evidence of the presence of this oxidised debris is provided in the high O/C atomic ratio particularly for GO, which cannot be explained solely by the presence of oxygen functionalities covalently bound to the CNMs.

Later on for GO, the ID/IG ratio began to decrease noticeably (Day 3) and the spectra started to lose the characteristic Raman signature of GO (Days 5 and 7). This was expected and is confirmation of degradation. As the disorder in the graphene-based sheet increased, an initial rise in the ID/IG ratio was observed after 1 day as expected due to the increased scattering of charge carriers by defects and as the crystalline graphitic domains become smaller35,37. This however will occur only up to a certain defect density, above which the material would start to be described as nano-crystalline. Any more defects would result in an 2 amorphisation of the sp carbon structure, resulting in a reduction in the ID/IG ratio and subsequently an attenuation of all characteristic peaks, which is indicative of an increasing 3 37 sp amorphous carbon phase . This change in the ID/IG as observed, corresponds to the predictions of the TK equation. The equation relates the size of sp2 crystal domains within a graphitic material as it undergoes degradation. The equation predicts that as the size of the 2 sp crystal domain decreases due to degradative influences, the ID/IG will increase until a certain minimum size, at which point the assumptions of the TK equation no longer apply. 38 The ID/IG will therefore start to decrease, as is observed . The equation has recently been adapted to describe increasing defect density in 2D graphene sheets induced by ion bombardment and confirmed with correlative scanning tunnelling microscopy analysis37. After Day 5, the D and G bands are no longer discernible and are unable to be identified at

Day 7. For this reason an ID/IG ratio has not been plotted. Rather a single low intensity peak exists which resembles the profile of a highly amorphous sp3 rich amorphous material38.

130

A similar phenomenon (reduction in the ID/IG ratio) was seen for OxNHs after day 5, which could be explained in the same manner. However OxNHs maintained their Raman spectrum fingerprints even after 7 days, hence confirming that some materials preserved some degree of their crystalline structure. In addition, the D’ peak at 1620 cm-1 was increasingly apparent over time for this material. This appearance is further evidence of defect formation within the graphitic backbone (Figure S8)39.

For OxMWNT, the D’ peak was already present in the material’s Raman spectra at day 0, and is a characteristic fingerprint of these oxidised/defected materials. However it was seen to increase with time, hence suggesting further degradation39. The trend of increasing then decreasing of the ID/IG ratio was less obvious relative to that seen for GO and OxNH. A pattern of increased (day 0 - 1 - 5 - 7) and then decreased (day 1 - 5) gradient was seen (Figure 3C), possibility due to the induction of nano-crystalline and amorphous carbon phases as consecutive graphene walls are exposed to the ClO-. Previous studies conducted by us, when attempting to demonstrate the biodegradation of functionalised MWNT both in 40 41 vivo and in vitro also showed an increase in the ID/IG ratios. The result was found when functionalised MWNT were internalised within microglia in the central nervous system. Others have also made similar observations29,42. These findings could be explained by the same mechanism, by which the outermost wall of the nanotubes gradually becomes dominated by amorphous structures following various oxidative degradation processes acting on it, resulting in the nanotubes becoming gradually less intact.

Morphological analysis using TEM and AFM confirmed the structural degradation of all tested materials and showed that patterns of degradation were specific to each material (Figure 4). For GO, the planar morphology was evident only up to day 2. Interestingly, after this time an increase in the thickness was seen, most likely due to the increased presence of oxidation reaction products that adsorb to the GO basal plane following incubation in NaClO18. After day 2, the GO sheets were seen to contain holes and defects indicative of changes in the nano-crystalline state, as suggested by the Raman data. Following this, the material became more amorphous with an unspecific morphology/ 3D structure which was eventually lost. These observations are similar to that reported previously in which the 17 horseradish peroxidase enzyme was incubated with H2O2 and GO . The enzyme was able to bind to GO and induce structural defects by catalysing the conversion of H2O2 to stronger oxidative species. For OxMWNT and as expected, the degradation pattern was different with visible exfoliation of the OxMWNT outer walls at day 5 and 7 (Figure S5). In our previous study on biodegradation of functionalised MWNT, a similar observation was seen in the brains of animals40. Many other studies have observed similar TEM image features when

131

MWNT are exposed to an oxidative influence43,44. But few have presented detailed imaging of the ultrastructural changes. For OxNH, the initial dahlia-like structure underwent a gross destruction, suggesting that oxidative degradation led to an increase of hydrophilicity of the individual nanohorns. Individualised nanohorns are seen at day 5 (Figure 4). After this point, there was an increase in amorphous materials amongst remaining OxNH crystals as suggested by the Raman spectra (Figure 3). Overall, the data suggest that NaClO is a substance with great capacity to degrade oxidised carbon nanomaterials via further oxidation.

The reasoning of why OxMWNT degraded at a slower rate than OxNH, is likely due to the intrinsic nature of the starting materials. Despite both being composed of rolled graphitic sheets with a hollow centre, the inner core of the NH (conical) is larger than in MWNT (tubular). This means that chemical reactions could more readily occur also from within the hollow centre of OxNH. In contrast, the inner hollow core of MWNTs is narrower and entering molecules such as ClO- would be more constrained by the physical environment (i.e. the internal diameter of the innermost nanotube). In addition, the conical structure of OxNH is more irregular with opening of holes occurring on the side walls45, allowing more accessibility for ClO- molecules and degradation to proceed. This is however not the case for OxMWNT, which maintain a more inherently crystalline non-defected structure within the inner walls. Finally, the closed end of one OxNH is more constrained (pentagonal arrangements) and so more reactive45 than the tips of one OxMWNT (no caps, only edges). These factors are due to the unique geometry of the 1D carbon nanostructures and therefore are not relevant to GO, which exists as high surface area, planar 2D sheets.

In addition to the aspect of morphology, the identity of the prevailing functional groups for each carbon nanostructure may be an important factor in driving the patterns and kinetics of degradation. This is due to the different relative stabilities and reactivities of epoxide, carbonyl and carboxylic acid functional groups towards nucleophilic agents, such as ClO-. For example, epoxide groups present in GO are particularly unstable and likely to react, due to the inductive effect of the oxygen and the unfavourable ring strain46; this may contribute to the observed hole formation (Figure 4) in the GO sheets16-18 where epoxides arise. Interestingly, ClO- can form epoxides from alkene double bonds47. This could be a mechanism leading to the initial oxidation of the un-oxidised inner walls of OxMWNT, which will then go on to be further oxidised as time goes by, leading to exfoliation (Figure S5). Hydroxyl and carbonyl groups can also react with ClO-23,48,49. However carboxylic acid groups – the furthest state of oxidation of oxygen functionalities – are somewhat stabilised due to the delocalisation of the π electron across the heteroatom. Nonetheless, carboxylic

132 acid groups can react via oxidative decarboxylation reactions50,51, essentially removing carbon and oxygen to form carbonates. None the less, the environment in which any functional groups are found can influence both their stability and reactivity.

Degradation of carbon nanomaterials via oxidative processes was further supported in our work by the dramatic changes between day 0 and day 7 in the FT-IR spectra for all nanomaterials incubated in NaClO 1 % (Figure S9). After 7 days in NaClO, it was observed that the assignment bands related to C-O (1200 – 1000 cm-1), C=O (~1726 cm-1) and O-H vibrations (3500 – 3000 cm-1) increased in intensity. Moreover, there was an increase in the aliphatic C-H stretch assignment band (2900 – 2800 cm-1) suggesting an increase sp3 C defective sites within the materials. These observations are in line with oxidative degradation. The 7 day spectra were however most likely a resulting from a combination of spectra from materials still undergoing degradation and of spectra from an array of degradation by-products. The visually more translucent appearance of the material suspensions at day 7 (Figure 2A) was suggestive of the well-dispersed and the increasingly non-graphitic nature of the by-products29. The detailed mechanism of the molecular degradation process should however be further investigated to isolate and characterise these degradation by-products.

Conclusion In this study, ClO-, from NaClO, has been shown to completely degrade GO 2D sheets with rapid kinetics compared to 1D OxMWNT and OxNH, which did not occur in H2O. Furthermore, detailed imaging of the morphological patterns by which oxidised carbon nanomaterials degrade has been provided. A summary schematic of the timeline and morphological degradation patterns for the various oxidised carbon nanomaterials tested is proposed below (Figure 5). Further studies are however required to isolate, identify and assess the environmental and biological impacts of the degradation by-products generated by ClO- upon reaction with oxidised carbon nanomaterials. As highlighted in other studies18, this is of particular importance since some of these degradation by-products could have chemical structures close to poly aromatic hydrocarbons, which are well-known carcinogenic molecules. Those experiments will help to complete the toxicological profile of these unique materials.

133

OxMWNT

OxNH

DayDay 0 0 Day 1 Day 3 Day 5 Day 7Day 7 NaClO 1%

Figure 4.5 Schematic representation of the degradation processes.

The progressive decay in the structural integrity of (A) GO, (B) OxMWNT and (C) OxNH are shown over time following incubation in 1 % NaClO.

134

4.6. Experimental.

Preparation of GO.

The oxidation of graphite was based on a modified Hummer’s method15. Briefly, a mixture of

0.8 g of Brazilian flake graphite (Graflake 9580) and 0.4 g of sodium nitrate (NaNO3) was maintained at approximately 0 ⁰C using an ice bath. A volume of 18.4 mL 96 % sulphuric acid (H2SO4) was added to the mixture while stirring. Then 2.4 g of potassium permanganate

(KMnO4) was added. The mixture was removed from the ice bath. After 30 min, the mixture’s viscosity increased, which resulted in the formation of a dark coloured paste. MilliQ H2O (37 mL) was added. The temperature was maintained at 98 ⁰C for 30 min. The mixture was subsequently diluted with 112 mL warm MilliQ H2O and 12 mL of H2O2 (30 %) of was added.

The suspension was centrifuged repeatedly at 9000 rpm for 20 min with MilliQ H2O, the process was repeated until the acidic pH of the supernatant was neutralised. At the intersection of the pellet and the supernatant, a single to few layer GO sheet containing orange-yellow gel like layer appeared. The supernatant was decanted and separated from the pellet and the orange-yellow gel like layer was carefully extracted with warm water. The extracted layer was then freeze dried for 48 hours (the samples was then weighed to determine the weight in a known volume of solution), reconstituted in a known volumes of MilliQ water and subject to 5 minutes sonication in a bath sonicator (VWR, UK). The obtained product was the GO material that was to be used in our experimentation. 10 mL of this product was lyophilised and the dried material weighed, allowing the mass concentration in the orginal aqueously dispersed product to be known. The concentration of the aqueously

GO product was found to be 2.4 mg/mL and was then adjusted to 2 mg/mL with MilliQ H2O. Samples were characterised via FT-IR, UV-vis, TEM, AFM, Raman spectroscopy and X-Ray photoelectron spectroscopy (XPS).

Preparation of OxMWNT.

Pristine MWNT (1g) (Nanoamor, USA) was sonicated in an ultrasonic cleaner (20 W, 40kHz) (VWR, UK) for 24 h in 150 mL sulphuric acid: nitric acid (3:1) 98 % and 65 %, respectively at room temperature. Ice was added to reduce the temperature. Water was added to the reaction mixture to allow for a 10-fold dilution. The diluted mixture was filtered through a 0.45 µm Omnipore membrane filter (Merk, UK) and the OxMWNT were suspended in MilliQ water, this was repeated until the pH was neutral. Samples were characterised as for GO.

135

Preparation of OxNH.

OxNH were prepared as previously described30. Briefly, single walled nanohorns (SWNH) were generated via laser evaporation of pure graphite (95%). SWNH (40 mg) of was dispersed in 50 mL 30 % H2O2. The dispersion was stirred on a hot plate set to 100 ˚C and irradiated with a Xe lamp ( =250 – 2000 nm, 3 W, 1 cm) allowing light-assisted oxidation.

The dispersion was filtered and washed multiple times with dH2O. The resulting samples were characterised as for GO.

Incubation of oxidised carbon nanomaterials with NaClO.

Dispersions of oxidised carbon nanomaterials were made up at 250 µg/mL with 1 % NaClO (Sigma-Aldrich, UK) in glass vials to 3 mL and gently stirred using a magnetic stirrer at room temperature (pH 10), pH was monitored daily to ensure maintenance. Each reaction was completed thrice. At days 0, 1, 2, 3, 5 and 7 after incubation 160 µL were removed from each vessel, transferred to an Eppendorf tube and made up to 660 µL with MilliQ H2O. The dispersion underwent centrifugation at 4000 rpm and the supernatant was discarded and replaced with 500 µL MilliQ water. This was repeated until the pH of the supernatant was neutral. This extracted material was then further analysed to monitor degradation. As controls, we used carbon nanomaterial dispersions (mass concentration: 250 µg/mL) in

MilliQ H2O, MilliQ H2O alone and 1 % NaClO. These were in an identical manner.

Thermogravimetric analysis.

Dried CNMs were placed in a crucible set at 25 ˚C in a TGA 400 thermogravimetric analyser (PerkinElmer, UK). The temperature was increased at 10 ˚C / min from 25 ˚C – 995 ˚C under a nitrogen flow of 20 mL/min.

X-Ray photoelectron spectroscopy.

X-ray Photoelectron Spectroscopy was conducted (Thermo Scientiﬁc, UK) at the NEXUS facility (the UK's National EPSRC XPS Users' Service, hosted by NanoLAB in Newcastle- upon-Tyne). XPS was accomplished using a Thermo Theta Probe XPS spectrometer (ThermoScientific, USA) with a monochromatic Al K-α source of 1486.68 eV. The survey XPS spectra were acquired with pass energy (PE) of 200 eV, 1 eV step size, 50 ms dwell time and averaged over 5 scans, etching time was set to 90 sec. The high resolution C1s XPS spectra were acquired with PE of 40 eV, step size of 0.1 eV, dwell time of 100 ms and averaged over 20 scans. Spectra of insulating samples were charge corrected via shifting the peaks to the adventitious carbon C 1s spectral component binding energy set to 284.6 eV and a linear background was subtracted. Assignments were made in accordance to

136

NIST’s XPS databases: π−π*: 290−290.2, O−C=O: 289.7−288.2 eV, C=O: 288.1−287.4 eV, C-O: 286.8– 286.2, C−C and C=C: 284.6 eV. CasaXPS (Casa Software LTD, UK) remote software was used to analyse the spectra, all data are reported in terms of atomic percentages.

UV-vis spectroscopy.

Samples (1 mL) were analysed in 1.5 mL quartz cuvettes using a Varian Cary 50 biospectrometer (Agilent Technologies, UK) and scanned between wavelengths of 200 - 800 nm.

Raman spectroscopy.

Spectra were collected using a micro-Raman spectrometer (Thermo scientific, UK) with a laser, λ = 633 nm that was calibrated using a polystyrene reference material. The calibrated instrument had an energy resolution of 2.5 cm-1. Spectra were collected at a laser power of 0.4 mW (GO) and 0.8 mW (OxMWNT and OxNH) at a magnification of 50 X (numerical aperture 0.75) with 25 sec exposure time, averaged over 30 locations. Spectra were considered between 250 - 3500 cm-1, the spectral resolution was 2.5 cm-1. The sample processing involved an auto-fluorescence 6th order polynomial baseline subtraction and normalisation of the scattering intensity according to the G band scattering intensity where present using OriginPro 8.5.1 software (Origin Lab, USA).

Transmission electron microscopy.

Aliquots of the respective dispersions were transferred to 400 mesh copper grids with a support film of carbon (CF400-Cu) (EMS, UK). Excess dispersion was removed with filter paper. The specimens were examined using a FEI Tecnai T-12 BioTWIN TEM (FEI, Eindhoven, NL), using 100 kV electron beam. Images were captured by an AMT digital camera (Gatan, UK).

Atomic force microscopy.

A multimode atomic force microscope (Bruker, UK) was applied in tapping mode. Using an Otespa tapping mode tip (Bruker, UK), scans were performed using the following parameters: a scan rate of 1 Hz, 512 lines per scan, an integral gain of 1 and a proportional gain of 5. Images were taken at 50 μm, 20 μm, 10 μm, 5 μm and 2 μm (1:1 aspect ratio). Post image processing was completed using Bruker Nanoscope Analysis Version 1.4 (Bruker, UK).

137

Attenuated total reflectance Fourier transform infrared spectroscopy.

Measurements were performed using a Tensor 27 FTIR machine (Bruker, UK) with a 3000 Series TM High Stability Temperature Controller with RS232 Control (Specac, UK) at 60°C. Data processing was completed using OriginPro 8.5.1 software (Origin Lab, USA)

4.7. Acknowledgements.

The authors would like to thank the Engineering and Physical Sciences Research Council NowNanoDTC program at the University of Manchester, UK (grants: EP/K016946/1 and EP/M010619/1) and the FP7 EC Graphene Flagship project (FP7-ICT-2013-FET-F604391) for financially supporting this research. We thank Prof Nicola Tirelli for kindly allowing the use of the FTIR. We thank the staff in the Faculty of Biology, Medicine and Health EM Facility, Dr Aleksandr Mironov and Ms Samantha Forbes, for their expertise and the Wellcome Trust for equipment grant support to the Facility. The University of Manchester Bioimaging Facility microscopes used in this study were purchased with grants from the BBSRC, Wellcome Trust and the University of Manchester Strategic Fund. We thank Dr Nigel Hodson for his expert advice in the use of atomic force microscopy and the NEXUS facility for the XPS service.

138

4.8. Supplementary information. A

System Na 1s O 1s C 1s S 2p Si 2p Purity O:C D Graphite - 3.0% 96.3% - 0.7% 99.3% 0.0 GO - 29.1% 70.9% - - 100.0% 0.4 MWNT - 1.1% 98.9% - - 100.0% 0.0 Ox-MWNT 0.3% 9.8% 89.7% 0.3% - 99.5% 0.1 NH - 2.6% 97.4% - - 100.0% 0.0 Ox-NH - 5.2% 94.8% - - 100.0% 0.1 Figure 4.S1 Analysis of functionalities present on the surface of starting oxidised carbon nanomaterials using XPS.

(A) The XPS survey spectra, where oxidised materials are shown in orange and (B) high resolution C 1s XPS spectra is shown for graphite (left panel), MWNT (middle panel) and NH (right panel) and (C) GO (left panel), OxMWNT (middle panel) and OxNH (right panel). (D) Based on the survey, the quantification of Na 1s, O 1s, C 1s, S 2p and Si 2p regions of GO, OxMWNT and OxNH has been tabulated. The spectra demonstrate that all materials have a relatively high purity with no more than 0.7% impurities. From the high resolution C 1s XPS spectra of GO, OxMWNT and OxNH and the corresponding starting materials, the percentage presence of π-π*, carboxylic groups (O-C=O), carbonyls (C=O), epoxides (C- O-C), hydroxyls (C-OH) and graphitic structure (C-C & C=C) has been determined (Figure 4.1D). All percentages represent atomic percentages. Spectra were acquired at the NEXUS facility and analysed by Dr Neus Lozano.

139

GO OxMWNT OxNH 100 100 100

80 80 80

60 60 60

40 40 40

% Weight % % Weight % 20 Weight % 20 20

0 0 0

200 400 600 800 1000% / loss weight Percentage 200 400 600 800 1000 200 400 600 800 1000 o o Temperature / C Temperature / C Temperature / C

Figure 4.S2 Chemical composition of the starting oxidised carbon nanomaterials as analysed by TGA.

Functionalisation of GO was found to be 39%, while for OxMWNT and OxNH the functionalisation degree was 30% for both OxMWNT and OxNH.

140

Day 0 Day 2 1.0 1.0 Day 1 1.0

0.5 0.5 0.5

Absorbance / a.u. Absorbance /

Absorbance / a.u. Absorbance / Absorbance / a.u. Absorbance /

0.0 0.0 0.0 400 500 600 700 800 400 500 600 700 800 400 500 600 700 800 Wavelength / nm Wavelength / nm Wavelength / nm

Day 3 Day 5 Day 7 1.0 1.0 1.0

0.5 0.5 0.5

Absorbance / a.u. Absorbance /

Absorbance / a.u. Absorbance / Absorbance / a.u. Absorbance /

0.0 0.0 0.0 400 500 600 700 800 400 500 600 700 800 400 500 600 700 800 Wavelength / nm Wavelength / nm Wavelength / nm

H2O GO H2O OxMWNT H2O OxNH H2O NaClO 1 % GO NaClO 1 % OxMWNT NaClO 1 % OxNH NaClO 1 %

Figure 4.S3 Evolution of the UV-vis absorption spectra of oxidised carbon nanomaterials in 1

% NaClO and H2O over time.

UV-vis absorption spectra was considered between λ = 400 - 800 nm to demonstrate the change in scattering and absorption of photons by the carbon nanomaterials over time in 1 % NaClO and H2O. The initial mass concentration at Day 0 was 0.25 mg/mL.

141

GO OxMWNT OxNH

Day 0 Day

Day1

Day2

Day3

Day5 Day7

Figure 4.S4 Representative observations of ultra-structural changes of oxidised carbon nanomaterials incubated in H2O as detected by TEM over time.

Samples of GO, OxMWNT and OxNH were removed and treated in the same way as 1% NaClO samples via centrifugation at Days 0, 1, 2, 3, 5 and 7, and analysed via TEM using a 100 kV beam (scale bars = 100 nm).

142

Figure 4.S5 TEM images of the wall exfoliation in OxMWNT when incubated in 1 % NaClO.

Images were taken after Day 5 of incubation in 1 % NaClO (scale bars = 50 nm).

143

A Day 0 Day 1 Day 2 2.5 2.5 2.5

2.0 2.0 2.0

ID/IG = 1.36 + 0.03 1.5 ID/IG= 1.36 + 0.03 1.5 1.5 ID/IG = 1.36 0.03 ID/IGn=30 = 1.36 0.03 ID/IG = 1.35 0.03 ID/IG= 1.35 + 0.03 1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Normalised Intensity Raman Normalised Intensity Raman Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 -1 -1 Raman Shift / cm Raman Shift / cm-1 Raman shift / cm Day 3 Day 5 Day 7 2.5 2.5 2.5

2.0 2.0 2.0

1.5 I /I = 1.36 + 0.03 1.5 D G I / I = 1.34 + 0.03 1.5 I /I = 1.33 + 0.03 ID/IG = 1.36 0.03 ID/IG D= 1.34G 0.03 IDD/IGG = 1.33 0.03

1.0 1.0 1.0

0.5 0.5 0.5

0.0 Normalised Intensity Raman 0.0

Normalised Intensity Raman 0.0

500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman -1 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm Raman Shift / cm-1 Raman Shift / cm-1

B Day 0 Day 1 Day 2 2.5 2.5 2.5

2.0 2.0 2.0

1.5 I /I = 1.38 + 0.05 I D/IG = 1.38 0.05 1.5 I /I = 1.39 0.03 1.5 I /I = 1.41 0.03 D G D G DID/IGG= 1.41 + 0.06 ID/IG= 1.39 + 0.03 1.0 1.0 1.0

0.5 0.5 0.5

0.0 Normalised Intensity Raman 0.0 0.0

500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman -1 Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm -1 Raman Shift / cm-1 Raman Shift / cm Day 3 Day 5 Day 7 2.5 2.5 2.5

2.0 2.0 2.0

I /I = 1.04 + 0.03 1.5 I /I = 1.3 0.21 1.5 D G D G ID/IG = 1.04 0.03 1.5 ID/IG= 1.16 + 0.3 ID/IG= 1.3 + 0.21 1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0

Normalised Intensity Raman 0.0 Normalised Intensity Raman

500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman -1 -1 500 1000 1500 2000 2500 3000 3500 Raman Shift/ cm Raman Shift/ cm -1 Raman Shift / cm

Figure 4.S6 Evolution of Raman spectra of GO over time.

Individual averaged Raman spectra are presented for GO when incubated in (A) H2O and (B) 1 % NaClO at time points Day 0, 1, 2, 3, 5 and 7 are shown.

144

A Day 0 Day 1 Day 2 2.5 2.5 2.5

2.0 2.0 2.0 I /I = 1.79 + 0.02 IDD/I G = 1.79 0.02 ID/IG = 1.75 0.03 ID/IG = 1.75 0.04 I /I = 1.75 + 0.03 D G 1.5 1.5 1.5 ID/IG= 1.75 + 0.04

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Normalised Intensity Raman Normalised Intensity Raman Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 -1 Raman Shift / cm-1 Raman Shift / cm Raman Shift / cm-1 Day 3 Day 5 Day 7 2.5 2.5 2.5

2.0 2.0 2.0 ID/IG = 1.77 0.03 ID/IG = 1.78 0.05 I /I = 1.73 0.03 1.5 I /I = 1.78 + 0.05 1.5 D G D G 1.5 I /I = 1.73 + 0.03 ID/IG= 1.77 + 0.03 D G

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Normalised Intensity Raman NormalisedIntensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm-1 Raman Shift/ cm-1 Raman Shift / cm-1 B Day 0 Day 1 Day 2

1.1 1.1 1.1 1.0 2.5 2.5 1.0 2.5 1.0 0.9 0.9 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 2.0 0.6 0.6 0.5 2.0 2.0 Normalised Intensity Raman 1500 1550 1600 1650 -1 0.5 Raman Shift / cm Normalised Intensity Raman 1500 1550 1600 1650 0.5 Normalised Intensity Raman 1500 1550 1600 1650 Raman Shift/ cm-1 Raman Shift / cm-1 1.5 1.5 1.5 I /II / =I =1.96 1.96 + 0.050.05 ID/IG= 1.97 + 0.04 IDID/I/IG= = 1.871.87 + 0.02 0.02 D GD G ID/IG = 1.97 0.04

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Normalised Intensity Raman Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 -1 Raman Shift / cm Raman Shift/ cm-1 Raman Shift / cm-1 Day 3 Day 5 Day 7

1.1 1.1 1.1 2.5 1.0 1.0 2.5 1.0 0.9 0.9 2.5 0.9 0.8 0.8 0.8 0.7 0.7 2.0 0.7 0.6 2.0 0.6 2.0 0.6

Normalised Intensity Raman 0.5 0.5 1500 1550 1600 1650 Normalised Intensity Raman 1500 1550 1600 1650 0.5 -1 Normalised Intensity Raman -1 Raman Shift / cm 1500 1550 1600 1650 Raman Shift / cm Raman Shift / cm-1 1.5 I /I = 2.00 + 0.04 D G 1.5 I /I = 2.01 + 0.06 1.5 ID/IG = 2.00 0.04 ID/IDGG = 2.01 0.06 I /I = 2.13 0.05 D GID /IG= 2.13 + 0.05 1.0 1.0 1.0

0.5 0.5 0.5

Normalised Intensity Raman 0.0 0.0 0.0 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman

500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman -1 -1 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm Raman Shift / cm -1 Raman Shift / cm

Figure 4.S7 Evolution of Raman spectra of OxMWNT over time.

Individual averaged Raman spectra are presented for OxMWNT when incubated in (A) H2O and (B) 1 % NaClO at time points Day 0, 1, 2, 3, 5 and 7 are shown. A zoom up of each spectrum when OxMWNT was incubated in 1 % NaClO over time is shown to demonstrate the development of D’ peak (arrow) as a further evidence of degradation.

145

A Day 0 Day 1 Day 2 2.5 2.5 2.5

2.0 2.0 2.0

ID/IG= 2.09 + 0.07 I /I = 2.12 0.05 1.5 ID/IG = 2.09 0.07 1.5 D IGD/I G=2.12 + 0.05 1.5 ID/IG = 2.04 0.07 ID/IG= 2.04 + 0.07

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Normalised Intensity Raman Normalised Intensity Raman Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm-1 Raman Shift / cm-1 Raman Shift / cm-1 Day 3 Day 5 Day 7 2.5 2.5 2.5

2.0 2.0 2.0

1.5 ID/IG = 2.05 0.07 1.5 ID/IG = 2.02 0.09 ID/IGI =/I 1.93= 1.93 + 0.050.05 ID/IG= 2.05 + 0.07 1.5 D G ID/IG= 2.02 + 0.09

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0

Normalised Intensity Raman 0.0

500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman -1 500 1000 1500 2000 2500 3000 3500 Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 -1 Raman Shift / cm Raman Shift cm Raman Shift / cm-1 B Day 0 Day 1 Day 2

1.1 1.1 1.1 1.0 1.0 2.5 1.0 2.5 0.9 0.9 2.5 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 2.0 0.6 2.0 0.6 0.5

2.0 NormalisedIntensity Raman 1500 1550 1600 1650 0.5

0.5 Raman Shift / cm-1 Normalised Intensity Raman 1500 1550 1600 1650 Normalised Intensity Raman 1500 1550 1600 1650 Raman Shift / cm-1 Raman Shift / cm-1 I /I = 2.03 + 0.12 1.5 1.5 1.5 D G I /I = 2.00 + 0.11 I /I = 2.22 + 0.10 I /I = 2.03 0.12 D G ID/IDG G= 2.02 0.10 D G ID/IG = 2.00 0.11 1.0 1.0 1.0

0.5 0.5 0.5

0.0

Normalised Intensity Raman 0.0 0.0 Normalised Intensity Raman 500 1000 1500 2000 2500 3000 3500 NormalisedIntensity Raman 500 1000 1500 2000 2500 3000 3500 500 1000 1500 2000 2500 3000 3500 Raman Shift / cm-1 -1 Raman Shift / cm Raman Shift / cm-1 Day 3 Day 5 Day 7

1.1 1.1 1.1 2.5 1.0 2.5 1.0 2.5 1.0 0.9 0.9 0.9

0.8 0.8 0.8

0.7 0.7 0.7 2.0 0.6 2.0 0.6 2.0 0.6

0.5 0.5 0.5

Normalised Intensity Raman Normalised Intensity Raman 1500 1550 1600 1650 1700 1500 1550 1600 1650 Normalised Intensity Raman 1500 1550 1600 1650 -1 Raman Shift / cm Raman Shift / cm-1 Raman Shift / cm-1

I /I = 2.13 + 0.09 ID/IG= 2.00 + 0.38 I /I = 1.42 + 0.28 1.5 D G 1.5 1.5 D G ID/IG = 2.13 0.09 ID/IG = 2.00 0.38 ID/IG = 1.42 0.28

1.0 1.0 1.0

0.5 0.5 0.5

0.0 0.0 0.0

Figure 4.S8 Evolution of Raman spectra of OxNH over time.

Individual averaged Raman spectra are presented for OxNH when incubated in (A) H2O and (A) 1 % NaClO at time points Days 0, 1, 2, 3, 5 and 7 are shown. A zoom-in of each spectrum when OxNH was incubated in 1 % NaClO over time is shown to demonstrate the development of D’ peak (arrow) as a further evidence of degradation.

146

Day 0 H2O Day 7 H2O GO Day 0 NaClO Day 7 NaClO

Absorbance (a.u.) Absorbance 4000 3000 2000 1000 Wavenumber / cm-1

OxMWNT Absorbance (a.u.) Absorbance 4000 3000 2000 1000 Wavenumber / cm-1

4000 3000 2000 1000

OxNH Absorbance / a.u. / Absorbance

4000 3000 2000 1000 Wavenumber / cm-1

Figure 4.S9 Evolution of ATR FTIR spectra of oxidised carbon nanomaterials over time.

Individual FTIR spectra (presented as absorbance) of the oxidised CNMs following incubation in H2O (Days 0 and 7) and 1 % NaClO (Days 0 and 7).

147

4.9. References

1 Novoselov, K. S. et al. Electric field effect in atomically thin carbon films. Science 306, 666-669 (2004). 2 Lee, C., Wei, X., Kysar, J. W. & Hone, J. Measurement of the elastic properties and intrinsic strength of monolayer graphene. Science 321, 385-388 (2008). 3 Balandin, A. A. Thermal properties of graphene and nanostructured carbon materials. Nat. Mater. 10, 569-581 (2011). 4 Falkovsky, L. A. Optical properties of graphene. J. Phys. Conf. Ser. 012004 (2008). 5 Park, S. & Ruoff, R. S. Chemical methods for the production of graphenes. Nat. Nanotechnol. 4, 217-224 (2009). 6 Sun, X. et al. Nano-graphene oxide for cellular imaging and drug delivery. Nano. Res. 1, 203-212 (2008). 7 Yoon, H. J. et al. Sensitive capture of circulating tumour cells by functionalized graphene oxide nanosheets. Nat. Nanotechnol. 8, 735-741 (2013). 8 Robinson, J. T. et al. Ultrasmall reduced graphene oxide with high near-infrared absorbance for photothermal therapy. J. Am. Chem. Soc. 133, 6825-6831 (2011). 9 Liu, S. et al. Antibacterial activity of graphite, graphite oxide, graphene oxide, and reduced graphene oxide: membrane and oxidative stress. ACS Nano 5, 6971-6980 (2011). 10 Nishida, E. et al. Graphene oxide coating facilitates the bioactivity of scaffold material for tissue engineering. Jpn. J. Appl. Phys. 53, 06JD04 (2014). 11 Su, Y. et al. Impermeable barrier films and protective coatings based on reduced graphene oxide. Nat. Commun. 5, 4843 (2014). 12 Abraham, J. et al. Tunable sieving of ions using graphene oxide membranes. Nat Nano 12, 546-550 (2017). 13 Kostarelos, K. & Novoselov, K. S. Materials science. Exploring the interface of graphene and biology. Science 344, 261-263 (2014). 14 Li, Y. et al. Graphene microsheets enter cells through spontaneous membrane penetration at edge asperities and corner sites. Proc. Natl. Acad. Sci. U.S.A. 110, 12295-12300 (2013). 15 Ali-Boucetta, H. et al. Graphene Oxide: Purified graphene oxide dispersions lack in vitro cytotoxicity and in vivo pathogenicity. Adv. Healthcare Mater. 2, 512-512 (2013). 16 Lalwani, G., Xing, W. & Sitharaman, B. Enzymatic degradation of oxidized and reduced graphene nanoribbons by lignin peroxidase. J. Mater. Chem. B 2, 6354- 6362 (2014).

148

17 Kotchey, G. P. et al. The enzymatic oxidation of graphene oxide. ACS Nano 5, 2098- 2108 (2011). 18 Bai, H. et al. Insight into the Mechanism of graphene oxide degradation via the photo-Fenton reaction. J. Phys. Chem. C 118, 10519-10529 (2014). 19 Kotchey, G. P. et al. A natural vanishing act: the enzyme-catalyzed degradation of carbon nanomaterials. Acc. Chem. Res. 45, 1770-1781 (2012). 20 Bussy, C., Bianco, A., Prato, M. & Kostarelos, K. Handbook of Safety Assessment of Nanomaterials: Toxicological Testing to Personalized Medicine Ch 11 (Pan Stanford Pub. Ltd., Stanford, 2014). 21 Zimmerman, G. & Strong, F. C. Equilibria and spectra of aqueous chlorine solutions. J. Am. Chem. Soc. 79, 2063-2066 (1957). 22 Winter, J., Ilbert, M., Graf, P. C. F., Özcelik, D. & Jakob, U. Bleach activates a redox- regulated chaperone by oxidative protein unfolding. Cell 135, 691-701 (2008). 23 Mirafzal, G. A. & Lozeva, A. M. Phase transfer catalyzed oxidation of alcohols with sodium hypochlorite. Tetrahedron Lett. 39, 7263-7266 (1998). 24 Klebanoff, S. J. Myeloperoxidase: friend and foe. J. Leukoc. Biol. 77, 598-625 (2005). 25 Malle, E., Marsche, G., Arnhold, J. & Davies, M. J. Modification of low-density lipoprotein by myeloperoxidase-derived oxidants and reagent hypochlorous acid. Biochim. Biophys. Acta. 1761, 392-415 (2006). 26 Hawkins, C. L. & Davies, M. J. Inactivation of protease inhibitors and lysozyme by hypochlorous acid: role of side-chain oxidation and protein unfolding in loss of biological function. Chem. Res. Toxicol. 18, 1600-1610 (2005). 27 Vlasova, II, Sokolov, A. V., Chekanov, A. V., Kostevich, V. A. & Vasil'ev, V. B. Myeloperoxidase-induced biodegradation of single-walled carbon nanotubes is mediated by hypochlorite. Bioorg. Khim. 37, 510-521 (2011). 28 Vlasova, II et al. PEGylated single-walled carbon nanotubes activate neutrophils to increase production of hypochlorous acid, the oxidant capable of degrading nanotubes. Toxicol. Appl. Pharmacol. 264, 131-142 (2012). 29 Kagan, V. E. et al. Carbon nanotubes degraded by neutrophil myeloperoxidase induce less pulmonary inflammation. Nat. Nanotechnol. 5, 354-359 (2010). 30 Lu, N., Li, J., Tian, R. & Peng, Y. Y. Binding of human serum albumin to single-walled carbon nanotubes activated neutrophils to increase production of hypochlorous acid, the oxidant capable of degrading nanotubes. Chem. Res. Toxicol. 27, 1070-1077 (2014). 31 Gorham, J. M. et al. Detecting carbon in carbon: exploiting differential charging to obtain information on the chemical identity and spatial location of carbon nanotube

149

aggregates in composites by imaging x-ray photoelectron spectroscopy. Carbon 96, 1208-1216 (2016). 32 Charsley, E. L. & Warrington, S. B. Compositional Analysis by Thermogravimetry Ch. 2 (Astm Intl, Baltimore, 1988). 33 Zhang, M., Yudasaka, M., Ajima, K., Miyawaki, J. & Iijima, S. Light-assisted oxidation of single-wall carbon nanohorns for abundant creation of oxygenated groups that enable chemical modifications with proteins to enhance biocompatibility. ACS Nano 1, 265-272 (2007). 34 Texter, J. Graphene dispersions. Curr. Opin. Colloid Interface Sci. 19, 163-174 (2014). 35 Thomsen, C. & Reich, S. Double Resonant Raman Scattering in Graphite. Phys. Rev. Lett. 85, 5214-5217 (2000). 36 Childres, I., Jaureguib, L. A., Parkb, W., Caoa, H. & Yong P. Chena. New Developments in Photon and Materials Research Ch. 19 (Nova Science Pub. Inc., New York, 2013). 37 Lucchese, M. M. et al. Quantifying ion-induced defects and Raman relaxation length in graphene. Carbon 48, 1592-1597 (2010). 38 Ferrari, A. C. Raman spectroscopy of graphene and graphite: Disorder, electron– phonon coupling, doping and nonadiabatic effects. Solid State Commun. 143, 47-57, (2007). 39 Dresselhaus, M. S., Jorio, A., Hofmann, M., Dresselhaus, G. & Saito, R. Perspectives on carbon nanotubes and graphene raman spectroscopy. Nano Letters 10, 751-758 (2010). 40 Nunes, A. et al. In vivo degradation of functionalized carbon nanotubes after stereotactic administration in the brain cortex. Nanomedicine (Lond). 7, 1485-1494 (2012). 41 Bussy, C., Hadad, C., Prato, M., Bianco, A. & Kostarelos, K. Intracellular degradation of chemically functionalized carbon nanotubes using a long-term primary microglial culture model. Nanoscale 8, 590-601 (2016). 42 Elgrabli, D. et al. Intracellular degradation of functionalized carbon nanotube/iron oxide hybrids is modulated by iron via nrf2 pathway. Sci. Rep. 7, 40997 (2017). 43 Russier, J. et al. Oxidative biodegradation of single - and multi-walled carbon nanotubes. Nanoscale 3, 893-896 (2011). 44 Liu, X., Hurt, R. H. & Kane, A. B. Biodurability of single-walled carbon nanotubes depends on surface functionalization. Carbon 48, 1961-1969 (2010).

150

45 Yudasaka, M., Iijima, S. & Crespi, V. H. Carbon Nanotubes: Advanced Topics in the Synthesis, Structure, Properties and Applications Ch. 19 (Springer - Verlag, Berlin, 2008). 46 Morgan, K. M. et al. Thermochemical Studies of Epoxides and Related Compounds. J. Org. Chem. 78, 4303-4311 (2013). 47 de Poorter, B. & Meunier, B. Catalytic epoxidation of aliphatic terminal olefins with sodium hypochlorite. Tetrahedron Lett. 25, 1895-1896 (1984). 48 Khurana, J. M., Sharma, P., Gogia, A. & Kandpal, B. M. Oxidative cleavage of 1, 2- diols, α-ketols and 1, 2-diketones with aqueous sodium hypochlorite. Org. Prep. Proc. Int. 39, 185-189 (2007). 49 Leduc, A. B. & Jamison, T. F. Continuous flow oxidation of alcohols and aldehydes utilizing bleach and catalytic tetrabutylammonium bromide. Org. Process Res. Dev. 16, 1082-1089 (2012). 50 Whistler, R. L. & Yagi, K. Further Application of the hypochlorite method of chain shortening in the carbohydrate series1. J. Org. Chem 26, 1050-1052 (1961). 51 Elmore, P. R. et al. Hypochlorite-induced oxidative decarboxylation of trisubstituted acetic acids. J. Org. Chem 54, 970-972 (1989).

151

Chapter 5

5. Splenic Capture and In Vivo Intracellular Biodegradation of Thin, Biological-grade Graphene Oxide Sheets

Leon Newman1,^, Dhifaf Jasim1,^, Eric Prestat2, Neus Lozano1, Irene De Lazaro1, Yein Nam1, Bakri M. Assas3,4, Sarah Haigh2, Cyrill Bussy1,*, Kostas Kostarelos1,*

1 Nanomedicine Lab, Faculty of Biology, Medicine & Health and National Graphene Institute, AV Hill Building, The University of Manchester, Manchester, UK. 2 School of Materials, The University of Manchester, Manchester M13 9PL, UK 3 Translational Medicine, Faculty of Biology, Medicine & Health, The University of Manchester, Manchester, UK 4 Department of Immunology, Faculty of Applied Sciences, King Abdulaziz University, Jeddah, Saudi Arabia

5.1. Statement.

Leon Newman, Dr Dhifaf Jasim, Dr Cyrill Bussy and Prof Kostas Kostarelos planned the study, designed the experiments. Leon Newman performed the materials characterisation and experiments regarding the long term fate and biodegradation of graphene oxide in the spleen, while Dr Dhifaf Jasim performed all the dose escalation and related toxicology studies, and provided the associated result descriptions and figures as stated. Leon Newman wrote the manuscript. Dr Eric Prestat acquired the electron diffraction patterns. The XPS data was generated at the NEXUS facility at the University of Newcastle and analysed by Dr Neus Lozano. Dr Neus Lozano also synthesised the graphene oxide. Irene De Lazaro generated and analysed the RT-qPCR data. All authors have critically reviewed and provided feedback on the manuscript. This manuscript has recently been submitted to Nature Nanotechnology.

Key Words: 2D materials, Biodegradation, Toxicology, Toxicity, Macrophage,

Nanomedicine______

^These authors contributed equally to this work * Correspondence to either: [email protected]

[email protected]

152

5.2. Abstract.

The development of nanomaterials for biomedical technologies is an area of research that is expected to yield medical breakthroughs. Carbon nanomaterials in particular, including 2D graphene-based materials (GBM) have shown promising applicability to drug delivery, tissue engineering, diagnostics and various biomedical strategies. However, to reap the benefits of the biomedical implementation of GBMs it is necessary to understand their possible toxicological implications and uncertain fate in vivo. In the literature, the toxicological profile and fate of GBMs is still controversial with contradictory outcomes from one investigation to the other. We had previously demonstrated that following intravenous administration, the 2D graphene oxide (GO) nanosheets are largely excreted via the urine; however a small but significant portion of the material remains sequestered within the spleen. Herein, we interrogate the potential toxicological consequences of this accumulation and the fate of the spleen-accumulated GO over a period of 9 months in C57BL/6 mice. We show that GO materials are not associated with any detectable pathological consequences for the spleen. Using confocal Raman mapping, we then determine the sub-organ biodistribution of GO at various time points after administration. The cellular identity and intracellular localisation of the material is also confirmed using immunohistochemistry coupled with Raman spectroscopy and transmission electron microscopy (TEM), respectively. This unique combination of techniques has identified cells of the splenic marginal zone as the main site of GO bioaccumulation. In addition, through analyses using both bright field -TEM coupled with electron diffraction, and Raman spectroscopy we reveal the fate of the spleen- accumulated GO over 9 months, and directly image the intracellular biodegradability of GO with ultrastructural precision. Our work offers highly encouraging and necessary results for the further development and exploitation of GO in biomedicine.

153

5.3. Introduction.

Graphene-based materials (GBMs) have the potential for exciting and revolutionary technological applications including those related to medical technologies such as improved drug delivery1-3. A key concern however is the implication of the interactions of these engineered nanomaterials with elements of the mammalian biological milieu4. In particular, understanding the fate of GBMs after their administration to the body and their tendency to accumulate within organs and cells of the mononuclear phagocytic system (MPS) such as the lungs, liver and spleen and the effects thereof5,6 are key to defining whether GBMs are suitable for biomedical applications or not.

The preferential accumulation of these materials in one element of the MPS over another is a function of the route of administration and the materials physicochemical characteristics7. We and others have previously reported the accumulation of functionalised nanosized GBMs in the spleen following both intravenous (i.v.)8 and intraperitoneal administration6. This splenic accumulation following i.v. administration is not restrictive only to GBMs5 and has been reported many times for various other nanomaterials9,10.

The spleen is a highly specialised organ and forms a major part of the MPS, uniting the innate and adaptive immune system11. It is specifically adapted to filtering blood via the strategic placement of an array of phagocytic cells such as specialised subpopulations of macrophages, which act to ‘clean’ the blood of foreign invaders and unwanted particulates. The organ is well known to be capable of the efficient removal of various materials including engineered nanomaterials. The filtered blood, free of the particulate matter is then returned to the main circulation via the efferent splenic vein11. The spleen also contains large numbers of lymphoid cells and antigen presenting cells such as dendritic cells. If the nanoparticles are recognised as antigens, these cells could induce an antigen specific immune response in the organism12,13. In addition to this, it has important roles in maintenance a healthy red blood cell population as well in the storage of iron11.

The accumulation of GBMs within the spleen, for example in macrophages, has been shown to be without toxicological consequences6. At the same time however, various other reports have detailed extensive immune responses characterised by the up regulation of inflammatory markers as well as histopathological changes5 following the administration of GBMs. These toxicological based discrepancies can often be linked to an inhomogeneity in the GBMs physicochemical properties such as thickness, lateral dimensions and the levels of surface functionalisation between investigations7,14. These variations can crucially

154 determine the implications of the materials on biological systems7. GO materials are often complex and poorly characterised and, this may explain the unclear state of the literature.

In addition to the toxicological influences of GBMs on the body, the physicochemical characteristics of GBMs can have potent implications on the long term fate and biodegradation of the materials4,15,16. The literature is populated with reports of the complete degradability of GBMs as well as other carbon nanomaterials under in vitro conditions mediated by oxidising chemicals17-19 as well as reactive oxygen and nitrogen species generating enzymes, including some that are relevant to mammals15,16,20-23. Other studies have also suggested that the colloidal attributes of GBM dispersions and the material parameters that can determine this have important implications for the materials biodegradability15. Low levels of oxidative functionalisation and increased thickness that can promote poor dispersibility as well as aggregation/agglomerates, are material properties that can hinder the enzyme mediated degradation of GBMs, while the contrary can promote it15,16. A study investigating the in vivo biodegradability of carboxylated graphene over a period of 3 months using confocal Raman spectroscopy confirmed the importance of these colloidal attributes, and reported incomplete degradation of graphene-based aggregates in different organs of Swiss Albino mice with associated pathological features5. In addition to this the visual changes to the individual flakes as degradation proceeded were not demonstrated. Moreover, the position of the carboxylated graphene was not shown with regards to the overall architecture of the organs.

In line with a recent comment on the need for more long term assessments concerning the interactions of nanostructures with mammalian organisms24, we describe a long term and detailed study characterising the in vivo safety as well as biodegradability profiles of graphene oxide (GO) nanosheets. GO nanosheets are a highly promising class of GBMs that are currently being intensively investigated for their applicability to various areas of biomedicine, where the safety and biodegradability of the materials are often imperative to their success.

We first interrogated the toxicological effects of a thoroughly characterised batch of GO nanosheets following their intravenous administration at different concentrations (2.5 to 10 mg/kg), on the structure and function of the murine spleen. Both haematological and immunological splenic functions were assessed for up to 1 month, in addition to the subjects overall health and weight. Following this, we describe a further series of experiments using an optimised treatment dose of GO, to determine the long-term cellular and intracellular biodistribution of GO within splenic architecture. Confocal Raman mapping in combination

155 with immunohistochemistry and fluorescence activated cell sorting (FACS) were applied, allowing the non-destructive mapping of the location of GO within the spleen. Finally using Raman spectroscopy, we track the biodegradability of GO in whole tissues, while in parallel we use bright field (BF) transmission electron microscopy (TEM) coupled with electron diffraction to directly probe the materials structural integrity within ultrathin tissue sections. Using this strategy, we reveal the splenic intracellular location of GO and its gradual biodegradative structural decay (both visually and spectroscopically) over a 9 month period as catalysed by the in vivo biological machinery.

5.4. Results.

We began our study with a necessary thorough characterisation of the synthesised GO nanosheets (Figure 5.1). The AFM and TEM micrographs evidence the synthesis of a population of single to few layer GO sheets with lateral dimensions predominantly below 2 µm that displayed a log normal distrubtion (Figure 5.1A and B). The Raman spectra provide evidence of the presence of an enhanced scattering peak in the D band relative to the G band (Figure 5.1C), indicative of the presence of defects. This has been further corroborated by the hexagonal diffraction pattern, which shows a degree of diffusivity within the diffraction rings, suggestive of the presence of crystalline and non-crystalline regions within the GO sheets (Figure 5.1D); the corresponding electron micrograph of the particular GO flake is shown in Figure 5.S2. ATR FT-IR, TGA and XPS (Figures 6.S1A, B, Ci and Cii) analyses indicate that these defected regions are occupied by oxygen containing species such as epoxide, hydroxyl, carbonyl and carboxylic acid groups. These defected regions constitute about 40% by weight of the total GO. UV-vis (Figure S1D) and fluorescence spectroscopies (Figure S1E) further revealed characteristic absorption and fluorescence curves expected for GO.

156

5 nm A

200 250

200 150 150 100 100

50 50 Count (n = (n Count 650) 0 = (n Count 650) 0 0 2 4 6 8 0 200 400 600 1 μm Thickness (nm) Lateral Dimension (nm) 0 nm B C D

I(D) / I(G) = 1.43 + 0.02

500 nm 1000 2000 3000 10 1/nm Raman Intensity (a.u.) Intensity Raman Raman Shift (cm-1)

Figure 5.1 Physicochemical characterisation of the starting GO material.

(A) AFM image, with corresponding graphs showing the distribution of sheet thickness and lateral dimensions. (B) TEM. (C) Raman spectroscopy with the I(D)/(G) ratio provided in the inset. (D) Electron diffraction pattern. Further physicochemical characterisation is shown in Figure 5.S1, with the SAED aperture micrograph (obtained by Dr Eric Prestat) of the corresponding GO sheet given in Figure 5.S2.

The effect of our characterised GO on the spleen structure following a single i.v. administration was then determined. The structure of the spleen was examined using haematoxylin and eosin (H & E) staining. The results revealed no damage or histopathological changes in the tissue neither after 24 h (Figure 5.2A) nor after 1 month (Figure 5.2B) post GO administration at doses of 2.5, 5 or 10 mg/kg. Magnified regions of both the red pulp and white pulp showed no apparent changes compared to those of mice injected with Dex 5% (negative control). Mice injected with 5 mg/kg lipopolysaccharide (LPS) (positive control), however, exhibited a reduction in the extent of haematoxylin staining, especially in the red pulp, which is indicative of cellular death.

It was observed that with increasing doses of administered GO, there was an increase in the number of cells showing brown pigmentation, especially in the red pulp and marginal zone regions of the spleens 24 h after the injection. At 1 month after the injection, the increase in the number of brown-pigmented cells was more prominent in these regions and the appearance of those brown-pigmented cells in the white pulp also became evident at the

157 maximum dosage (Figure 5.2C). In addition to the microarchitecture, the gross appearance of the mice spleens was examined (Figure 5.S3). At the early stages, i.e. 24 h after the injection, there was a notable darkening of the spleens especially following the injection of the highest GO doses. Upon staining the spleens with terminal deoxynucleotidyl transferase (TdT) dUTP Nick-End Labeling (TUNEL) stain, we evidenced a significantly higher number of TUNEL positive apoptotic cells at the highest dose injected (10 mg/kg), shown in Figures 6.S4 and 6.S5. The latter figure represents the quantification of TUNEL positive cells per mm2 area.

158

A C 24 h 160 24h *** 1 month 5% 120 LPS at 24h * *

Dex * 80 *

pigmentation 40 Cells Cells withbrown

0 Dex 5% LPS 5 GO 2.5 GO 5 GO 10

mg/ml mg/kg mg/kg mg/kg

GO2.5 mg/kg

GO5 mg/kg

GOmg/kg10 LPSmg/kg5

B 1 month

Dex

GO2.5 mg/kg

GO5 mg/kg GOmg/kg10

Figure 5.2 Effect of GO on spleen structure.

H&E stained spleen sections (5 µm thick) are shown after injection of GO at (2.5, 5 and 10 mg/kg), Dex 5% (negative control) and 5 mg/kg LPS (positive control) after (A) 24 h and (B) 1 month post injection. The first column is comprised of 20x magnification images. Scale bars represent 100 µm. The middle and last columns show 60x magnifications of red and white pulps, respectively. The scale bars represent 50 µm. The inset images show 300x magnified regions with scale bars that represent 5 µm. (C) Quantification of the total cells with brown pigmentation detected in spleen sections. Data are represented as mean ± standard error (SE). Statistical significance was assessed via the Kruskall – Wallis test with Dunns Post hoc test (p < 0.05 *, p < 0.005 *** vs. age matched negative control). Mice were n = 3-4 mice per group. Dr Dhifaf Jasim performed the experiments and contributed this Figure.

159

We decided next to extend our examination to interrogate the splenic function (haematological and immunological). The results describing the haematological function are provided in Figures 5.3A and 5.3B. Figure 5.3A presents the optical microscopy images of H & E stained blood smears of mice injected with GO at doses of 2.5, 5 or 10 mg/kg, after 24 h and 1 month compared to those of mice injected with Dex 5% (negative control) or 5 mg/kg LPS (positive control). The black arrows indicate deviant red blood cells (RBCs) in a LPS- injected mouse blood smear. The quantification of the deviant RBCs is demonstrated in the same figure. The graph provides evidence of normal levels of aberrant RBCs (not more than 3%25) for all groups injected with GO at the different concentrations, both at 24 h and 1 month after administration. The LPS-injected group (positive control) however demonstrated elevated levels of abnormal RBCs relative to the negative control (injected with Dex 5%) but this difference was not statistically significant.

Perls’ Prussian blue staining was used to detect the hemosiderin content of the mouse spleens at 24 h and 1 month after being injected with different concentrations of GO (2.5, 5 or 10 mg/kg) (Figure 5.3B). As expected, the levels of haemosiderin increased with the age of the mice and a significant reduction in the amount of haemosiderin was found for LPS- injected mice. At both 24 h and 1 month after injection and irrespective of the concentration used, no significant differences were observed between mice injected with GO compared to age-matched mice injected with Dex 5%.

160

24 h 1 month

Dex

LPS at 24h at LPS GO2.5 mg/kg

24h 10 1 month 8 * LPS at 24h

GO5 mg/kg 6 4 2

% Deviant% RBCs GOmg/kg10

B 24 h 1 month

Dex

LPS at 24h at LPS GO2.5 mg/kg

24h 1 month 3 LPS at 24h

2 GO5 mg/kg 1 *

% % Perls' Prussian blue GOmg/kg10

Figure 5.3 Effect of GO on spleen haematological function, after 24 h and 1 month of injection of GO at different concentrations (2.5, 5, 10 mg/ml) compared to control mice, Dex 5% (negative control) and LPS 5 mg/kg injected (positive control).

(A) Blood smears of mice stained with H & E at 2 magnifications 60x (scale bar 50 µm) and inset images 300x (scale bar 5 µm). The graph represents the percentage of deviant RBCs (reticulocytes, abnormal and aged RBCs) from the total RBC counts; the black arrows indicate some of the deviant RBCs detected from LPS-injected mice. The graph was obtained after counting > 500 cells from each mouse (n = 3 for each group). (B) Pearls’ Persian blue stained spleen tissue sections (5 µm thick) at 2 magnifications 20x (scale bar 100 µm) and inset images 100x (scale bar 20 µm). Blue spots demonstrate regions of Persian blue stained haemosiderin. The graph represents the percentage of total imaged tissue that stained blue indicating the presence of haemosiderin. This was obtained by measuring the intensity of the blue stained regions from 3-4 images from each mouse captured at 20x (n = 3-4 mice per group) by Fuji image J analysis.

161

The immunological function of the spleen was also investigated (Figure 5.4A and 5.4B). Representative T-cell lymphocyte counts for both the helper CD4+ and cytotoxic CD8+ T- lymphocytes are provided in Figure 5.4A. The graph shows there were slight changes in CD4+ or CD8+ cell populations in mice injected with different concentrations of GO (2.5, 5 and 10 mg/kg) after 24 h or 1 month compared to their corresponding age-matched negative controls that had been injected with Dex 5%. In contrast, the LPS-injected mice showed a significant reduction in such cell populations. Representative flow cytometry plots outlining CD4+ and CD8+ T-lymphocyte counts are displayed in Figure 5.S6. The gene expression profiles of commonly excreted inflammatory cytokines, including the pro-inflammatory (IL-6, IL-1β and TNF-α) and the anti-inflammatory (IL-10 and TGF-β) cytokines, were evaluated (Figure 5.4B). No significant differences in the levels of the pro-inflammatory cytokines IL-6 and TNF-α were detected in any group compared to mice injected with Dex 5%. Slight elevations in the levels of IL-1β however were detected in GO-injected mice at 24 h. After 1 month these levels returned to normal for the mice administered with the lower 2 concentrations of GO (2.5 and 5 mg/kg) but were found to be significantly reduced for the higher concentration (10 mg/kg). There was also a significant reduction in anti-inflammatory cytokines IL-10 and TGF-β, especially in mice injected with the highest dose (10 mg/kg) after 24 h and at all doses after 1 month. Based on this finding, the cell apoptosis detected and our previous studies26, an i.v. GO dose of 10 mg/kg was considered unsatisfactory for a long-term investigation (i.e. above 1 month). Therefore, a concentration of 7.5 mg/kg was selected to explore the long-term fate (up to 9 months) of the GO material within the spleen. At this concentration, the growth curve of mice was normal over the course of the 270 days (Figure 5.S7) and no symptoms of distress were observed.

162

A B i 30 24h 25 * 1 month LPS at 24h 6 20 * IL-6 15 *** 10 4

% % CD4 Cells 5 0 2 Dex 5% LPS 5 GO 2.5 GO 5 GO 10 mg/kg mg/kg mg/kg mg/kg 0

25 IL-1β 6 *** ii 20 15 * *** 4

10 * * inflammatory -

% % CD8 cells 5 2 Pro 0 *** Dex 5% LPS 5 GO 2.5 GO 5 GO 10 0 mg/kg mg/kg mg/kg mg/kg

6 TNF-α 4

6 Gene expression (relative to to (relative dextrose) expressionGene *** IL-10 4

2 * ***

0 inflammatory

6 - TGF-β 4 Anti TGF-β 2 6 * *** 04 2

Gene expressionGene

Dex 5%

(relative to dextrose)

GO 5mg/kg GO GO 5mg/kgGO

LPS 5mg/kg LPS

GO 10mg/kg GO 10mg/kg GO

GO 2.5mg/kg GO 2.5mg/kg GO 24h 1 month

Figure 5.4 Effect of GO on spleen immunological function, after 24 h and 1 month of injection of GO at different concentrations (2.5, 5, 10 mg/ml) compared to control mice, Dex 5% (negative control) and LPS 5 mg/kg injected (positive control).

A) Spleen T-lymphocyte cell counts, CD4+ T-helper cell counts plot (left) and cytotoxic CD8+ cell count plot (right). Cell counting was performed by flow cytometry. Three repeat counts were performed for each group of mouse (n = 3 mice). (B) Gene expression levels of pro-inflammatory cytokines IL-6, IL- 1β and TNF-α and gene expression levels of anti-inflammatory cytokines IL-10 and TGF-β. A and B represent the spleen haematological function, while C and D represent the spleen immunological functions. In A, B and C data are represented as mean ± standard error (SE), while the RT-q-PCR in D, data are represented as mean ± standard deviation (SD). Statistical significance was performed by the Kruskal- Wallis multiple comparison test and Dunns post hoc test . While the Welch ANOVA and Games-Howell’s post-hoc test was used in D. In all cases statistics are p < 0.05 *, p < 0.01 **, p < 0.005 *** vs. negative control.

163

As part of a long-term study to determine the fate of GO in the spleen, we decided to analyse the sub-splenic biodistribution of the GO over time. We administered a dose of GO (7.5 mg/kg) that was observed not to yield untoward weight deviations or symptoms of distress (Figure 5.S7), but allowed the facile detection of the material in spleen samples using Raman spectroscopy and TEM. Using confocal Raman mapping, the bioaccumulation pattern of GO within the spleen was revealed (Figure 5.5A) and a higher accumulation of GO in specific anatomical regions of the spleen was demonstrated. This suggested the possible preferential accumulation of GO in certain splenocyte sub-populations. Using immunofluorescence in correlation with conventional (point-and-shoot) Raman spectroscopy on spleen sections, we found that GO was primarily sequestered within a distinct subset of splenic macrophages. These were shown to be the marginal zone macrophages (MZM), which stained positive for the ERTR-9 antigen and that Raman analysis showed to contain GO. These cells are typically located on the border regions between the red and white pulp within the splenic marginal zone (Figure 4B). This preferential uptake was further confirmed by FACS in combination with conventional Raman spectroscopy on isolated cell sub populations (Figure 5.5C and Figure 5.S9).

In addition to the cellular preferential location of GO, we observed that over time, there was a loss of the cellular accumulated GO Raman spectral integrity (Figures 5.5A and 5.5B) and a corresponding decrease in the abundance of detectable GO positive cells in spleen sections (Figure 5.5C). While the GO Raman signature was easy to identify at 24 h, the same signature could hardly be detected at the later time points (Figures 5.5A and 5.5B). The greatest loss in GO positive cells was observed during the first month post GO administration but it also continued to decrease over the remaining course of the experiment (Figure 5.5A, 5.5B and 5.5C). The changes overtime suggested a loss of the spleen- accumulated material possibly via a biodegradation process. Apoptosis was excluded using the TUNEL assay, which showed no increase in the apoptosis in the regions of the spleen where GO was identified (Figures 5.5D and 5.S8).

164

A B 1

2 1 1

3 Day Day Day

1 2

3 Day 30 Day 1 30 Day

1 2

Day 270 Day

Day 270 Day 3

C D 60 * 200 GO

45 ) 2 150 Dex 5% 30

100

Cell / mm / Cell (

15 50

Postive for GO (%) GO for Postive

Percentage of MZM MZM of Percentage Apoptotic Cell Density Density Cell Apoptotic 0 0 1 30 180 270 Day 1 Day 30 Day 270 Time (Days) Figure 5.5 Splenic localisation and abundance of GO at different time points following the administration of 7.5 mg/kg GO.

(A) Confocal Raman maps of the GO Raman signature (orange-red) detected in mice spleen sections at different time points post GO administration (Day 1, 30 and 270). The length scale bars represents 200 µm while the Raman intensity bar represents 5 x 105 – 1 x 106 a.u between 1000 – 2000 cm-1 (B) An overlay image of the immunostaining of MZM (green), nuclei (blue) with bright field images, together with correlative Raman spectra of regions of interest, at different time points (Day 1, 30 and 270). The white circles and their corresponding numerical identities indicate different regions of interest where the Raman spectra were acquired. The corresponding numbered spectra are inset in the respective images. Scale bars represent 50 μm. (C) The percentage of cell sorted MZMs that screened Raman positive for GO over time (n = 20 spectra for each 3 animals at each time point). (D) Apoptotic cell density detected as per the TUNEL assay, for GO and Dex 5% treated groups at different time points (Day 1, 30 and 270). Statistical significances were tested using the Kruskal - Wallis multiple comparison test with Dunn’s Post hoc test. A full break down of the results of the statistical analysis for Graph 6.4C and 6.4D is given in Table 6.S1 and 6.S2, respectively. In all cases, statistics significant where p < 0.05.

165

In order to confirm whether the material underwent biodegradation within the spleen, we exercised an array of analytical techniques in an intricate and detailed manner as described in the experimental section. Using conventional Raman spectroscopy on physically homogenised spleen tissue, we reproducibly detected GO at different time points in the study that allowed us to probe GO biodegradation at the whole organ level. We saw again that the intensity of the Raman signature of GO decreased over time and by 9 months the GO signature was scarcely detectable (Figure 5.6Ai). The complete series of averaged spectra (n = 30) over time is shown in Figure 5.S10. We further scrutinised the peak scattering intensity ratio in the D vs. G bands (I(D)/ I(G)) over time (Figure 5.6C). The way in which this ratio evolves can reveal a wealth of information regarding the structural changes that are on-going27. We saw an initial overall increase in the distribution of the I(D)/ I(G) ratio during the first 2 weeks following administration. At the 1-month time point however, there was a significant decrease in the ratio. As time progressed, the ratio continued to decrease and the scattering peaks became progressively less pronounced. These changes were indicative of a gradual decay of the crystal structure of the material27.

We then attempted to visually probe the abundance of the materials at the cell level. Using H & E staining, we observed the unique increased presence of slightly darkened regions within the marginal zones of the spleen (Figure 5.6Aii). This corroborates the data in Figure 5.5A, 5.5B and 5.S9), which identified the MZMs as the principal cell responsible for GO accumulation. These dark regions were scarcely present in the group administered with Dex 5% (Figure 5.S11), but where present were attributed to iron-based haemosiderin. The number of dark regions that could be found in the marginal zone over time was quantified (Figure 5.6B). We saw that the number of dark regions within the marginal zone significantly decreased over time as seen from the steep negative gradient. In addition, the abundance of marginal zone specific dark regions approached values equivalent to those recorded for the Dex 5% treated mice by Day 270 (Figure 5.6B).

In order to directly observe and characterise the state of the materials within the ultrastructure of the splenic marginal zone cells, we optimised a TEM protocol as described in the experimental section. It was evident that the materials morphologically evolved over time within vesicular compartments of marginal zone cells (Figure 5.6Aiii and 5.S12). The GO sheets appeared to lose their characteristic sheet-like morphology, becoming increasingly less defined and harder to detect over time. By Day 30, there was evidence of a transition state, where some of the vesicle-bound GO sheets were shown to have developed holes throughout their sheet-like structure (Figure 5.S14). We further developed a protocol to qualitatively assess the extent of the structural damage and crystallinity of the GO using

166 electron diffraction. We observed a clear trend overtime with a decreasing ability of the vesicle entrapped GO to display an electron diffraction signature characteristic of GO (Figures 5.6Aiv and 5.S13). This confirmed the reduction in the structural integrity of the GO over time, already revealed by Raman spectroscopy (Figures 5.6A, 5.6C and 5.S10). The structural changes that we detail were not observed to a significant extent in the starting material when incubated at 37 ⁰C in darkness over 9 months in parallel to our in vivo work, which implicates the in vivo biology as the mediator of these changes (Figure 5.S15 and 5.S16).

167

A Day 1 Day 30 Day 270 i) .) a.u 600 600 600

400 400 400 200 200 200

0 0 0 1000 1500 2000 Raman Intensity Intensity ( Raman 1000 1500 2000 1000 1500 2000 -1 Raman Shift (cm-1) Raman Shift (cm-1) Raman Shift (cm ) ii)

iii)

) iv) 4 6 6 6

4 4 4

x10 Counts ( 2 2 2

0 0 0

4 6 8 10 4 6 8 10 4 6 8 10 Intensity Intensity Reciprocal Distance (1/nm) Reciprocal Distance (1/nm) Reciprocal Distance (1/nm) B GO Treated C 40 Dex 5% Treated 2.0 * 30 1.8 1.6

20 1.4

10 I(D)/I(G) 1.2

Dark areas / MZ / areas Dark 0 1.0 0 50 100 150 200 250 300 1 7 14 30 90 180 270 Time (Days) Time (Days) Figure 5.6 Overview of the splenic degradation of GO over 9 months following its intravenous administration (7.5 mg/kg).

(A) i) The average Raman spectra of GO present in physically homogenised spleen tissue at different time points, n=10 X 3 ii) Splenic sections of mice that were stained with H&E, scale bars represent 50 µm. Inset images shows the presence of GO material in the vicinity of cells of the marginal zone, scale bars represent 10 µm. iii) TEM micrographs of GO sequestered within the vesicular compartments of marginal zone splenocytes over time, scale bars represent 1 µm. The inset shows a magnification of the GO material at the respective time points, scale bars represent 500 nm. iv) Representative electron diffraction pattern line profiles of GO at different time points post administration, with the respective diffractogram inset, scale bars represent 10 1/nm. Dr Eric Prestat captured the electron diffraction patterns shown here. (B) The average number of dark areas present in the splenic marginal zone at each time point (n= 5 X 3) (p < 0.05 *). (C) The evolution of the Raman I(D)/I(G) ratio over 9 months (n=3X10). All statistical significances for Graph 5.5B and 5.5C are given in Table 5.S3 and 5.S4, respectively. A more detailed view of degradation is given in Figures 5.S8 to 5.S11.

168

5.5. Discussion.

Two criteria are preferable in the development of engineered nanomaterials for biomedical applications: a safe in vivo profile in the context of the designated application and a potential for biodegradation or overtime elimination4. Amongst GBMs, GO sheets are the most studied oxidised sp2 hybridised carbon nanomaterial derivatives for biomedical applications owing to their hydrophilic nature and the possibility to functionalise them easily, allowing them to be used as nano-carriers for the delivery of pharmaceutical agents2,28,29. However, in contrast to other more established biomedical carbon based nanovectors, information concerning the in vivo safety profile and fate overtime of GO is limited4. This is despite this information being of crucial importance for the translation of this exciting material from laboratory research to tangible biomedical applications. Following their i.v. administration, we have previously determined the spleen to be the site of accumulation of GO nanosheets that are not initially excreted via the kidney 8. The main aim of our study was therefore to follow up on this work and address the overtime in vivo toxicological implications of intravenously administered GO sheets and their local bioaccumulation in the spleen. We also sought to determine the in vivo biodegradability of GO sheets and therefore whether the accumulation in the spleen is a transient event or permanent.

An extensive physicochemical characterisation of our synthesised GO material was first performed (Figure 5.1). Various reports have detailed the key importance of properly defining and characterising the GBMs that are to be used in biomedical research4,14,30. Characteristics such as lateral dimensions, thickness and functionalisation are key determinants of a GBMs toxicological influence on biological systems7. We confirmed the successful synthesis of single and up to few layers GO sheets with lateral dimensions of up to 2 µm. The size distribution could be described by a log normal distribution. The population was clearly skewed towards the lower lateral dimensions (Figure 5.1A). Literature studies recommend when injecting colloidal particles intravenously that the materials dimensions do not exceed the diameter of a capillary (5 - 7 µm), as the particles must be able to journey through these narrow vessels following their administration without getting trapped, for example in the lungs7,31,32. Spectroscopic analysis was consistent with the successful functionalisation of the sp2 carbon hexagonal lattice with various interspersed oxygen functionalities (Figures 5.1C, 5.S1). The presence of polar oxygen containing functionalities minimises the thermodynamic consequences of hosting a hydrophobic carbon based material in the hydrogen bonding network of aqueous environments33. This ultimately allows the formulation of a well dispersed aqueous GO preparation. Overall, our characterisation results demonstrate the successful preparation of GO nanosheet dispersions that were

169 suitable for i.v. administration7,32. The classification framework for GBMs defined by the European Graphene Flagship Future and Emerging Technologies Project, considers the 3 physicochemical parameters we assessed: number of layers, lateral dimensions and level of oxygen rich functionalisation, to be fundamental for their biological consideration as they are likely to modulate GBM-organism interactions14. Our material is therefore suitable for biomedical research.

We had previously described the accumulation of retained GO in the spleen following i.v. administration8. This splenic bioaccumulation of GO is in good agreement with results from other independent laboratories6,34. Despite this, we could find no studies that examined the consequences of GO accumulation on the splenic function at these dosages. This prompted us to explore the effects of GO accumulation on the splenic structure and function.

Using H & E staining and Perls’ Prussian blue staining, we found no apparent histopathological damage or changes in splenic morphology at both 24 h and 1 month post administration compared to negative Dex 5% treated mice (Figures 5.2 and 5.3B). This is consistent with other literature data6, although some reports have demonstrated structural changes associated with significant pathological implications35. This is perhaps explicable by inter-laboratory differences in the materials physicochemical characteristics, especially thickness, lateral dimensions and dispersibility of the materials used7. When analysing the different major compartments of the spleen, we observed that the red pulp and white pulp areas were visually healthy in GO injected mice.

The 2 main functions of the spleen, namely haematological and immunological functions, were then interrogated. The haematological function concerns the ability of the red pulp macrophages to eliminate aberrant and old RBCs36-40. Aberrant RBCs normally do not account for more than 3% of total RBCs25. The assessment of blood smears for all mice that were treated with GO at the different doses (2.5, 5 and 10 mg/kg) both after 24 h and 1 month showed no rise in the number of deviant or abnormal RBCs (Figure 5.3A). This indicated that the splenic function was conserved allowing the maintenance of a healthy RBC population41. The other aspect of the haematological function is the capacity of the spleen to serve as an iron reservoir by storing the biologically important metal as haemosiderin in red pulp macrophages42-44. It was determined that there were no significant differences in the haemosiderin levels in all the mice injected with GO at different concentrations at 24 h and 1 month compared to the corresponding age-matched negative controls. This indicates that the spleen’s iron-storage capabilities were also preserved. Other studies have indicated no changes in the overall haematological parameters (complete blood count) of mice injected with single doses of GBM up to 10 mg/kg up to 48 h45 and 10 days46,

170 or even as high as 60 mg/kg after 24 h47. The same finding is true after repeated administration of 10 mg/kg every 2 days up to 22 days48, 15 mg/kg every other day for 14 days49 and repeated oral administration of very high doses of GO (60 mg/kg) every 24 h for 5 consecutive days50. Herein, the LPS (positive control) administered group had an increased population of defected RBCs (black arrows in Figure 5.3A) and showed elevated numbers (more than the expected physiological values > 3%25) compared to all other groups. This treatment group also displayed a significant reduction in the amount of hemosiderin in the red pulp. This could be due to the marked cell death induced by LPS51, that was not induced by GO, which would compromise the splenic haematological function.

Regarding the immunological function of the spleen, we first focused on the cell-mediated immunity which was measured by determining the populations of T-lymphocytes specifically CD4+ and CD8+ cells (among the major immune population in the spleen white pulp), as previously described 52,53. Cell-mediated immunity has a crucial role in the recognition and elimination of antigenic substances such as abnormal cells and foreign matter13. It was found that there were slight differences in both cell populations in mice injected with different concentrations of GO (2.5, 5 and 10 mg/kg) after 24 h and 1 month compared to the corresponding age-matched negative controls (Figure 5.4A). This is in agreement with the findings of Wang et al. who showed no changes in CD4+ and CD8+ T-cell populations in the spleens of C57BL/6 mice injected with 2 different types of graphene nanosheets at 1 mg/kg up to 7 days after i.v. injection54. The same study demonstrated there was a significant rise in such cellular populations after the administration of the same dose of multi-walled carbon nanotubes (MWNT) and that this elevation persisted after 7 days54. Such results illustrate how different morphologies of oxidised sp2 carbon nanomaterials may interact with biological systems in different ways and thereby resulting in different biological outcomes4.

To extend our study concerning the immunological function of the spleen, we analysed the gene expression profiles of various cytokines, including the pro-inflammatory (IL-6, IL-1β and TNF-α) and anti-inflammatory (IL-10 and TGF- β) cytokines (Figure 5.4B). The release of cytokines is evidence of CD4+ T cell activation, which is known to mediate CD8+ cells and B- cell functions12. No significant changes in the levels of pro-inflammatory cytokines IL-6 and TNF-α were detected in any group compared to those in mice injected with Dex 5%. This was in agreement with the findings of Wang et al. who also did not detect any changes in cytokine levels following the i.v. administration of graphene nanosheets even 1 week post administration54. In contrast, the rise in IL-1β levels that we detect in all GO-injected mice after 24 h was noted too by Orecchioni et al. after 24 h of incubation of GO with human immune cells55. The levels of IL-1β returned to normal values after 1 month (2.5 and 5

171 mg/kg), indicating that this effect is a transient mild inflammatory response occurring soon after injection. Moreover, after 24 hours post administration, it was observed that there was a suppression of IL-10 and TGF- β especially at the highest injected dose. In addition to this, after a month following the administration of GO at all tested doses, it was curiously noticed that these same two cytokines (IL-10 and TGF- β) were also suppressed. While IL-1β demonstrated significant suppression after 1 month following only the highest tested dose. Furthermore, at 1 month following the administration of GO at the highest dose, there was a significant suppression of pro-inflammatory cytokine IL-1β. A significant suppression of the anti-inflammatory cytokines IL-10 and TGF- β was also detected at this time point.

These suppressive influences of GO, that were not detected following the administration of Dex 5% have previously been reported by Sydlic et al. following the administration of GO at 20 mg/kg. The authors attributed this to a protective mechanism56. Since the pro- inflammatory cytokine levels were only slightly affected (except for IL-1β at 10mg/kg after 1 month), these results could suggest that the suppression may be due to direct inhibitory effects of GO on secreting cells (e.g. primarily macrophages in case of TGF-β57 or monocytes and macrophages in case of both IL-1058 and IL-1β59) by a mechanism that does not involve secondary regulatory effects60 or a decreased responsiveness of such cells61. T- cells are the main source of IL-662, while TNF-α is secreted by macrophages but it can as well be produced by many other cells such as CD4 T-lymphocytes. The T-cell populations were unaltered in this study, which could further explain the unchanged levels of the latter 2 cytokines. This may suggest the involvement of the innate immune system55 however more in detailed studies are required to explore this in greater depth.

In terms of the overall health, we regularly inspected and monitored the animals for symptoms of distress and followed their weight gain63. Normal weight gains and behaviours were observed for all mice following injection at all dosages with no obvious symptoms of distress. However, we have previously seen an uncharacteristic growth curve in mice injected with a single dose of GO at 10 mg/kg after 1 month26 which may indicate that this dose is approaching a toxicological threshold.

Additionally, the cytokine inhibitory effects at the same high dose and cellular apoptosis prompted us to determine the lowest dose that would allow for ease of GO detection without compromising the health of the animals using weight loss as criteria. In order to do this, we incrementally decreased the dose of administered GO and ensured there were no abnormalities in the growth curve compared to the Dex 5% treated mice (Figure 5.S7). It was concluded based on the measurements that a dose of 7.5 mg/kg was optimal.

172

During the analysis of the spleen H & E stained sections, a dark colouration of the tissue was evidenced, that increased with greater dosages of GO (Figure 5.2 and Figure 5.S3). This could be on account of the presence of GO. Wen et al. demonstrated a similar dark black colouration of the spleens of mice injected with functionalised GO sheets (fGO) that increased with an increasing concentration of fGO (up to 16 mg/kg)64. There was also an increase in the population of brown-pigmented cells in the marginal zone and red pulp of injected mice, suggesting the presence of GO in these particular cells. The observation was particularly apparent at the highest dose. In our previous study, a similar appearance of dark brown/black pigmented cells was noted in the red pulp of spleens of mice injected with fGO8. The pigmentation of the spleens in this previous study however, was darker and more obvious, perhaps due to the thicker nature of the fGO that was injected compared to the non-functionalised GO used herein. Other studies have also shown black local regions in the H & E stained tissue sections of organs (including lung and liver) upon administration of fGO45,54,65 and GO. This dark matter was attributed in these cases to the accumulation of foreign carbon nanostructures66.

Brown pigmentation in the spleen sections however, could be indicative of substances other than GO, such as iron that is known to be stored as haemosiderin preferentially in the splenic red pulp 67,68. Thus, Perls’ Prussian blue stain (which stains specifically haemosiderin blue) was therefore used to interrogate whether the observed dark material was the iron- based haemosiderin or GO. Most of the dark material in the red pulp splenocytes was confirmed to be haemosiderin. In contrast however, in or in close proximity to cells of the marginal zone of GO treated mice we observed the presence of a brown material (unstained by the Perls’ Prussian Blue dye) that was obviously present in the spleen of the GO treated groups relative to the control groups (Figures 5.3B). It was unambiguously confirmed that this brown material was GO using confocal Raman mapping (Figure 5.5A). This particular region of the spleen is densely packed with various strategically located phagocytic cells and is the site of accumulation of particulate antigens and apoptotic cells9. Literature studies have previously detailed the accumulation of other engineered nanoparticles in this region including lipid based nanoparticles9,69 and CNT34,70.

The next step was to determine the identity of the specific cells that sequestered GO within the splenic marginal zone. By combining Raman spectroscopy with both immunofluorescence (Figure 5.5B) and FACS (Figure 5.5C and 5.S9) independently, it was determined that the cell population responsible for the predominant GO splenic accumulation was a subpopulation of splenic macrophages, known as marginal zone macrophages (MZMs). These cells make up a small fraction of the total splenocytes and are situated throughout the middle and outer portion of the marginal zone, which separates the red and

173 white pulp71. In this region of the spleen, blood that was previously flowing under high pressure due to the restrictive diameter of the terminal arteriole emerges from the white pulp and into the sinuses of the marginal zone. Due to the open nature of this region, there is a lack of resistance which causes the blood to experience a dramatic deceleration in flow velocity and therefore resulting in a local decrease in blood pressure9,11. The highly phagocytic MZMs, take the opportunity provided by the decrease in blood pressure9,11 to remove particulate matter from the blood which includes cellular debris and exogenous particulate matter9,11. We infer that the GO sheets are endocytosed by the MZMs. Literature reports detail the accumulation of various nanoparticles including liposomes9,10, polystyrene microspheres9 and carbon nanotubes34 in these highly phagocytic cells. In the present study, other cells including metallophillic macrophages, red pulp macrophages and dendritic cells were also identified to uptake GO. However, the amount of material up taken by these other cells was disproportionately smaller compared to the amount of GO up taken by the MZMs (Figure 5.S9).

The residence of the GO within the marginal zone macrophages was then probed over time. After the first day, it became increasingly difficult to detect GO using confocal Raman mapping as well as conventional ‘point and shoot Raman spectroscopy (Figures 5.5A, B and C). The majority of the detected GO Raman spectra appeared to be increasingly attenuated and distorted with time and could even be described as predominantly featureless by Day 270. The main decrease was evident between Day 1 and Day 30 (Figure 5.5A, B and C). Results from confocal Raman mapping of spleen sections agreed with the conventional Raman spectroscopy. They showed a clear decrease of intensity of the GO mapped Raman signals between Day 1 and Day 30. This may be due to the biodegradation of the material, leaving GO crystalline regions of diminutive size that were below the limit of detection of conventional Raman spectroscopy72. Another possibility is the bioelimination of the material unaccompanied by significant biodegradation. It is important to note that the decrease in detectable GO was not due to apoptosis of the cells found to uptake GO as revealed by a TUNEL assay (Figure 5.5D). The apoptosis assay revealed a significant initial increase in cell death at Day 1 relative to the corresponding Dex 5% control group; however, the extent of cell death was still within the norm expected for spleen tissue73,74. By looking at the images, it is apparent that this initial detection of apoptotic cells occurred within the red pulp, and not in the marginal zone where the densest accumulation of GO was observed (Figure 5.S4 and Figure 5.S8). The extent of apoptosis then receded within 1 month to show insignificant differences to that of the Dex 5% group. This means that this phenomenon is transient.

174

In order to determine whether the accumulated GO completely biodegraded in vivo or had been eliminated, an intricate series of analytical experiments was optimised. Conventional Raman spectroscopy of GO present in homogenised tissues revealed that over 270 days following administration, the GO Raman signature followed the 3 stage classification of disorder of graphitic carbon crystals27(Figure 5.56i, C and 5.S10). This theory which describes the transformation of sp2 carbon structures into mainly tetrahedral sp3 amorphous carbon phases, is known as the amorphous trajectory of carbon27. The GO starting material before injection, displayed a prominent D peak (Figure 5.1C), which is due to the breathing modes of sp2 carbons specifically present in the structure’s rings27. In order for the D peak to be visualised, charge carriers must be inelastically scattered by a phonon and subsequently undergo a second elastic scattering by a defect within the material75. The presence of such a heightened D peak indicates that prior to administration, the material was significantly defective. This is expected considering the conditions of the modified Hummers’ method used to oxidise and exfoliate graphite to form GO76,77.

One week after administration, the I(D)/I(G) ratio of GO in the spleen began to rise, which indicated that the degree of defects within the materials had increased. The rise in the ratio however was slow. This could be due to the presence of oxidised debris physisorbed to the surface of the sheet which degrade first and offer some protection to the underlying GO sheet retarding the initial biodegradative process. The presence of this oxidised debris is alluded to by the XPS O/C atomic ratio of 0.5, which cannot be explained solely by the presence of graphene oxide oxygen functionalities alone. This protective effect may also be enhanced also by the in vivo derived protein corona which may also influence the evolution of the material. The increase in the I(D)/I(G) occurs for up to approximately 2 weeks within the spleen. The I(D)/I(G) ratio subsequently decreased, which signifies the second stage of the amorphous trajectory of carbon and can be understood as the gradual increase in defects within the 2D structure resulting in the increasing presence of amorphous carbon structures. The increasing amorphicity causes the overall attenuation of the GO Raman signature, which in turn induces a decreased intensity of peaks therefore a decrease in the I(D)/I(G) ratio78 (Figure 5.6C). Towards the final time point, it became difficult to discern the D and G bands, these spectra were therefore not included for further analysis of the evolution of I(D)/I(G) ratio over time. The pattern of an increasing and then decreasing I(D)/I(G) ratio is predicted by the Tuinstra-Koenin equation which has recently been adapted to 2D graphene sheets78. The equation relates the size of sp2 crystal domains within a graphenic material as it undergoes degradation. The equation predicts that as the size of the 2 sp crystal domain decreases the ID/IG will increase until a certain minimum size, at this point

175 the ID/IG will decrease as the peaks are gradually attenuated and the assumptions of the TK equation no longer hold.

This explains the increasingly difficult GO Raman signal acquisition after 1 month compared to 1 day post GO administration, irrespective of the Raman techniques used - conventional Raman spectroscopy (1 µm x 1 µm spot size) or confocal Raman mapping (large tissue section area, mm2 size). Moreover, the attenuation of the 2 characteristic Raman scattering peaks continues to occur with the consequent decrease in the I(D)/I(G) ratio until all peaks are attenuated, which is majorly observed with limited exceptions, by Day 270 (Figure 5.6Ai). These results strongly suggested that the accumulated GO was biodegrading in the spleen over a 9-month period. The same changes were not evident for the material incubated in water in vitro at body temperature, for the same amount of time (Figure 5.S15).

Despite the spectroscopic inferences, it was desirable to further visually confirm the biodegradation of the materials within the splenocytes using microscopic techniques. BF- TEM was therefore used to directly image the materials within the ultrastructure of cells in vivo (Figure 5.6Aiii). Over time, GO appeared to lose its characteristic crystalline structure and become increasingly disordered, as confirmed by the electron diffraction patterns of the intracellularly entrapped GO (Figure 5.6Aiv). An acceleration voltage of 80 kV was used, in order to reduce knock-on damage79. At Day 1, it was relatively easy to locate intracellular GO structures and acquire hexagonal diffraction patterns characteristic of GO within the cellular compartments. However at longer time points, there was an overall decrease of the abundance of material as well as of the intensity measured for the diffraction peaks associated with GO (Figures 5.6Aiv and 5.S13). This was attributed to a progressive attenuation of the crystallinity of the GO to a more amorphous structure due to biodegradation15. This conclusion is supported by the Raman analysis and BF-TEM data (Figure 5.6Ai and Aiii). At the 30 days’ time point, holes started to appear in the in the intracellular GO sheets (Figure 5.S14) that were not apparent in the starting material (Figure 5.1C, 5.S15 and 5.S16). We confirmed that this was not due to electron beam damage during image acquisition. These observations may suggest a biodegradation GO transition state, whereby defects accumulate within the material and then erode away, resulting in the creation of large observable holes within the GO sheets. Similar results have been documented several times in vitro when sp2 hybridised carbon nanomaterial and their oxidised derivatives, including GBMs, were exposed to peroxidase family enzymes15,16,20 as well as other enzymes such as cellular nicotinamide adenine dinucleotide phosphate- oxidase and induced nitrogen oxide synthase80. After 270 days, some material is still visually present within the presented intracellular compartments of the observed cells. The diffraction

176 intensity of these remaining materials is very weak and barely distinguishable from the tissue/ embedding resin amorphous background. These materials are likely to be mainly amorphous in nature and composite of sp2 and sp3 carbon based materials intermixed with the vesicular contents. These materials are likely to be highly oxidised and will be further degraded and cleared with time.

An interesting question following the observed biodegradation is the nature of the resulting by-products. Although we have not determined the molecular identity of these in vivo generated species, that maybe include oxidised aromatic species with lacking obvious crystalline structure, our study reveals that their production did not obviously jeopardise the splenic health. Previous studies have shown that the in vivo administration of carbon nanomaterial biodegradation by-products generated in vitro, lack toxic consequences for animal subjects81,82 in support of our findings. However, further studies are required to determine the components of the cell that catalyse the observed biodegradation as well as the molecular identity of the degradation by-products that result.

In conclusion, evidence has been provided to suggest that following their i.v. administration GO nanosheets that accumulate in the spleen do not induce histopathologic changes. Importantly, we report the maintenance of the functional (both haematological and immunological) integrity of the spleen at doses as high as 10 mg/kg, although at the highest dose administered slight immunological variations were observed. Using Raman mapping as well as immunohistochemistry and Raman in combination to map the splenic distribution of GO, we identified that a sub population of the splenic macrophages known as the MZMs are the predominant cells responsible for bioaccumulating the material. Finally using an array of techniques including Raman spectroscopy and TEM coupled with electron diffraction, the fate of the bioaccumulated GO was tracked over a 9-month period and provide compelling evidence to suggest the in vivo biodegradation of GO to such an extent that there is a complete loss of its crystalline structure. This work details the first time a GBM has been visually shown to undergo complete structural biodegradation down to the ultrastructural level with the complete disappearance of its graphitic crystallinity under in vivo conditions. Our study provides evidence to suggest that the initial splenic accumulation of GO in the spleen is a transient event. This important positive result supports the further development of thoroughly characterised GO nanosheets for biomedical technologies and is highly encouraging especially for drug delivery purposes, where the nanocarriers demonstrated safety and biodegradability is an advantage83.

177

5.6. Experimental.

Graphene oxide (GO) preparation.

The oxidation of graphite was based on a modified Hummers’ method68. Briefly, a mixture of

0.8 g of Brazilian flake graphite (Graflake 9580, Brazil) and 0.4 g of sodium nitrate (NaNO3) was maintained at approximately 0⁰C using an ice bath. Then 18.4 mL 96% sulphuric acid

(H2SO4) was added to the mixture while stirring, followed by the gentle addition of 2.4 g of potassium permanganate (KMnO4). The mixture was then taken from the ice bath. After 30 min, the mixture increased in viscosity and resulted in the formation of a dark coloured paste. A volume of 37 mL of MilliQ H2O was then added. An increase in temperature to 80⁰C was observed and this continued until 98˚C. The temperature was then maintained at 98⁰C for 30 min. After this duration, the mixture was diluted with a volume of 112 mL of warm

MilliQ H2O. A 12 mL aliquot of 30% H2O2 was subsequently added. The suspension was centrifuged and pelleted repeatedly at 9000 rpm for 20 min; the supernatant was then removed and replaced with MilliQ H2O. This process was repeated until the acidic pH of the supernatant was neutralised (pH 7, 20°C) and a viscous orange/ brown gel layer of single to few layer GO appeared on top of the graphite oxidation by-products. This gel-like layer was carefully extracted with warm water to yield a dispersion of GO. A known volume of this aqueously dispersed material was freeze dried and the dry material weighed. This allowed the concentration of the extracted GO material to be determined. The concentration of the material was found to be 2.4 mg/mL. The aqueously dispersed GO was then subjected to ultrasonic sonication for 5 min (VWR, 80W) to prepare the desired small sheets size GO. The batch reference of GO used in the present study is GO (f1s). Samples were characterised via atomic force microscopy (AFM), transmission electron microscopy (TEM), selected area electron diffraction (SAED), attenuated total reflectance Fourier transform infrared spectroscopy (ATR FT-IR), thermo gravimetric analysis (TGA), X-Ray photoelectron scattering (XPS), UV-vis and fluorescence spectroscopy to confirm successful synthesis of single to few layer small GO sheets.

Atomic force microscopy.

Sample preparation was completed on freshly cleaved mica, treated with 40 μL poly-L-

Lysine (Sigma-Aldrich, UK). The poly-L-Lysine was used to present a positively charged surface that would allow the adhesion of GO. Aliquots 10 μL of GO (100 µg/mL) were then transferred on to the mica-poly-L-lysine surface and left to adsorb for 2 min. Unbound structures were removed by gentle rinsing with 2 mL of MilliQ H2O and left to dry at room temperature. A multimode atomic force microscope (Bruker, UK) was then applied in tapping

178 mode, in order to avoid damaging the samples, for height (trace and retrace) and amplitude measurements. Measurements were performed using an Otespa tapping mode tip (Bruker, UK). The following parameters were employed: a scan rate of 1 Hz; lines per scan of 512; an integral gain of 1 and a proportional gain of 5; an amplitude set point value (150 mV) was maintained approximately constant between all measurements. Scans were taken at 50 μm, 20 μm, 10 μm, 5 μm and 2 μm (aspect ratio 1:1). Post image processing was completed using the Bruker Nanoscope Analysis Version 1.4 software (Bruker, UK) and included section analysis for measuring cross sectional height of samples.

Transmission electron microscopy for materials characterisation.

A 20 µL aliquot of sample (200 µg/mL) was placed on a carbon-coated copper grid (CF400- Cu) (Electron Microscopy Services, UK) and left to adsorb for 2 minutes. Filter paper (Merk- Millipore, UK) was used to gently remove the excess of dispersed material. Then samples were observed with a FEI Tecnai 12 BioTWIN microscope (FEI, NL) at an acceleration voltage of 100 kV. Images were taken with a Gatan Orius SC1000 CCD camera (GATAN, UK).

Raman spectroscopy for materials characterisation.

Samples were prepared for analysis by drop casting ~20 μL of GO (100 ug/mL) dispersion on to a glass slide. Samples were left to dry for at least 2 h at 37°C. Spectra were collected using a micro-Raman spectrometer (Thermo Scientific, UK) using a λ=633 nm laser. Spectra were considered between 250–3500 cm-1, enabling the visualisation of the D and G scatter peaks. Prior to the measurements, the Raman system was energy calibrated to a polystyrene reference material. The calibrated instrument had an energy resolution of 2.5cm- 1. Spectra were collected at a laser power of 0.4 mW with a magnification lens of 50 X (numerical aperture: 0.75) with 25 s exposure time, and averaged over 5 locations. Spectral processing involved a 6th order polynomial auto fluorescence baseline subtraction, all spectra were then normalised to the G peak for I(D)/I(G) calculation using Origin Pro 8.5.1.

ATR Fourier transform infrared spectroscopy.

A drop of GO dispersion was placed into a Tensor 27 FTIR machine (Bruker, UK) and dried for 5 min at 60°C until a powder remained. Spectroscopic analysis was carried out between 750 and 3500 cm-1. Data processing was completed using OriginPro 8.5.1 software (Origin Lab, USA).

179

Thermogravimetric analysis.

Lyophilised GO (2 mg) was weighed into a ceramic crucible set at 25 °C in a TGA 400 thermogravimetric analyser (TGA) (PerkinElmer, UK). The temperature was increased at a rate of 10°C per minute from 25°C to 995 °C, with a nitrogen flow of 20 mL/min. Measurements were considered between 100 – 800 ºC to avoid water interference. Data processing was completed using OriginPro 8.5.1 software (Origin Lab, USA).

X-Ray photoelectron spectroscopy.

The chemical composition of the GO sheets were studied by XPS at the NEXUS facility (the UK's National EPSRC XPS Users' Service, hosted by nanoLAB in Newcastle-upon-Tyne). XPS spectra was recorded using a Thermo Theta Probe XPS spectrometer with a monochromatic Al K-α source of 1486.68 eV. The survey XPS spectra were acquired with a pass energy (PE) of 200 eV, a 1 eV step size, a 50 ms dwell time and averaged over a total of 5 scans. The etching step was 90 s. The high resolution C1s XPS spectra were acquired with PE of 40 eV, a step size of 0.1 eV, a 100 ms dwell time and averaged over a total of 20 scans. Spectra from insulating samples were charge-corrected by shifting all peaks to the adventitious carbon C1s spectral component, with binding energy set to 284.6 eV. The CasaXPS software was used to process the spectra acquired at NEXUS. Processing steps included the C1s XPS spectral deconvolution and assignment of the different deconvoluted spectral components as per the binding energies at which they arise. These assignments were made in accordance to NIST’s XPS databases: π−π*: 290−290.2, O−C=O: 289.7−288.2 eV, C=O: 288.1−287.4 eV, C-O: 286.8– 286.2, C−C and C=C: 284.6 eV. The full width half maximum of each deconvoluted peak, other than that of the π – π* was constrained to the same extent and a Shirley background was taken. All percentages represent atomic percentages.

UV/Visible spectroscopy.

The UV-vis absorbance spectra of GO was measured by a Varian Cary win UV 50 Bio spectrophotometer (Agilent Technologies, UK). The GO samples were diluted in water from 7.5 to 20 µL/mL prior to measurement in a 1 mL glass cuvette with 1 cm path length. Dual- beam mode and baseline correction were used throughout the measurements to scan the peak wavelength and maximum absorbance between 200 and 600 nm.

180

Fluorescence spectroscopy.

Fluorescence emission spectra of GO were measured over a concentration range of between 75 to 200 µg/mL, using a LS-50B spectrofluorimeter (PerkinElmer, UK), with both excitation and emission slits set at 20 nm. The excitation wavelength used was 525 nm.

Animal handling procedures.

Eight-week-old C57BL/6 mice (18.1 ± 0.7 g) were obtained from Harlan (Oxfordshire, UK), were allowed to acclimatise for 1 week and were maintained under a 12 h light/dark cycle at a steady temperature and humidity. Mice were allowed access to food and water ad libitum for the duration of the investigation. All experiments were conducted in accordance with the UK Home Office Code of Practice (1989) for the housing and care of animals in scientific procedures in accordance with prior approval from the UK Home Office (Project licence number: P6C826A12). For short term splenic function and structural studies, mice were intravenously administered with GO (2.5, 5 and 10 mg/kg) or Dex 5% (negative control) and were culled for the different experiments after 24 h and 1 month. LPS (5 mg/kg) at 24 h was used as the positive control. These mice were only maintained for 24 h, since the mice showed symptoms of distress and were therefore culled at this time point. For longer term degradation studies mice were administered with an optimised dose of 7.5 mg/kg or Dex 5% (negative control). The mass concentrations of GO in the injection was 0.75 mg/mL, corresponding to 150 μg of GO per injection. Mice were divided into different experimental groups (n=3 for both GO and control Dex 5% treatment groups) based on the time points of analysis: 1, 7, 14, 30, 90, 180 and 270 days. Mice were culled at their respective time points via being subject to terminal anaesthesia followed by cervical dislocation. For all experiments mice were continuously inspected, monitored and weighed every 4 days during the first month and weekly thereafter.

Spleen histo-pathological analysis.

Spleens of mice were fixed with 4% paraformaldehyde and dehydrated through increasing concentrations of alcohol (ethanol 50-100%) and imbedded with molten paraffin. Sections of 5 µm were cut on a microtome (Leica, UK) and stained with haematoxylin and eosin stain (H & E); cover slipped with resinous mounting media and allowed to dry for 48 h at 37ºC. Images were collected using a 20x objective in a 3D Histech Panoramic 250 Flash slide scanner. Images were processed and analysed using Pannoramic Viewer (http://www.3dhistech.com/) and Fiji/ImageJ software (version 1.5c; National Institutes of Health, Bethesda, MD). At least 3 mice were examined per condition.

181

Blood smear analysis.

Blood was collected from mice via cardiac puncture and immediately drawn into a capillary tube. One drop of blood (approximately 4 mm in diameter) was placed on a clean slide (near the end). The drop was spread by another glass slide placed at a 45° angle and then pressed into the drop of blood allowing it to spread via capillary action. The slides were then fixed with cold methanol (precooled at -20°C) and then allowed to dry at room temperature. The samples were stained by H & E stain and imaged as described above. Reticulocytes and abnormal and aged RBCs were considered as deviant RBCs35-39,76, and their percentage of the total RBCs was calculated. More than 500 cells were counted for each mouse (n = 3 for each group).

Perls’ Persian blue analysis of spleen sections.

Paraffin-embedded sections of spleens were sectioned at 5 µm and then dehydrated as described above. The sections were then were stained with a freshly prepared mixture of aqueous solutions of 20% hydrochloric acid and 10% potassium ferrocyanide trihydrate (Sigma, UK) at a volume ratio of 1:1, followed by counter staining with nuclear fast red 0.2% aqueous solution. The slides were dehydrated through 95% and 100% ethanol solutions respectively, with 2 consecutive changes for each concentration. They were then cover slipped with resinous mounting media and left to dry overnight. Images were collected as described above. The percentage of blue pigmentation was analysed by Fiji/ImageJ software (version 1.5c; National Institutes of Health, Bethesda, MD). This was obtained from 3 or 4 images from each mouse captured at 20x after setting a threshold (n=3-4 mice per group).

Spleen single-cell suspensions preparation.

Half spleens were extracted from mice and placed immediately in RPMI 1640 media (Sigma, UK) supplemented with 10% FBS (Gibco, Thermo Scientific, UK) and 1% penicillin & streptomycin (Sigma, UK). Single-cell suspensions were obtained by gently passing the spleens through a 100 µm cell strainer (BD Falcon cell sieve). The cell suspension was centrifuged at 400 g for 5 min and the supernatant was discarded. The cells were then re- suspended in fresh media and contaminating RBCs were removed by adding 3–4 mL red cell lysis buffer (Roche, UK) and allowed to react at 4⁰C for 5 min. The cell suspension was then centrifuged (400 g for 5 min) and the cell pelleted while supernatant was discarded. The cells pellet was then re-suspended and washed in fresh media twice. Splenocytes were counted using a cytometer, aliquoted and further processed as needed.

182

T-lymphocyte counts by flow cytometry.

Splenocytes were first pre-incubated with anti-CD16/32 antibodies at 2 ug/mL (eBiosceince, UK, 16-0161-85) for 20 min on ice, in order to block non-specific binding sites. The cells were then washed (centrifuge at 400 g for 5 min) and then labelled with the antibody mixture of anti-CD4-PE at 0.25 µg/mL (eBioscience, UK, 12-0041-82) and anti-CD8a at 0.25 µg/mL. PerCP-Cy5.5 (eBioscience, UK, ref. #45-0081-82) was further incubated with the cells on ice for another 20 min. Cells were then washed and re-suspended in buffer (1% sodium azide and 1% FBS in 1X PBS) for acquisition. Flow cytometric analysis was carried out using a BD FACSVerse flow cytometer (Becton Dickinson, Oxford, UK). Single-colour control samples were used to set compensation. PE and PerCE-Cy5.5 fluorescence was detected with 574/26 nm and 690/50 nm band pass filters, respectively, after excitation with a 488 nm blue LASER. Data were analysed and statistics were generated using the FACSuite v1.6 software (Becton Dickinson, Oxford, UK).

Spleen cytokine gene expression by RT-qPCR.

An Aurum Total RNA Minikit (Biorad, UK) was used to extract the total RNA of 2 × 106 isolated splenocytes. The RNA concentration and quality were analysed using UV spectrophotometry (BioPhotometer, Eppendorf, UK). cDNA synthesis was performed from 1 µg RNA sample with iScript cDNA synthesis kit (Bio-Rad, UK) according to manufacturer’s instructions. The protocol for reverse transcription was set as follows: 25°C for 5 min, 42°C for 30 min, 85°C for 5 min and 4°C for 5 min. Two µL of cDNA sample were used for each real-time qPCR reaction performed with iQ SYBR Green Supermix (Bio-Rad, UK). Experimental duplicates of each sample were run on CFX-96 Real Time System (Bio-Rad, UK) with the following protocol: 95°C for 3 min, 1 cycle; 95°C for 10 s, 60°C for 30 s, – repeated for 40 cycles. Melt curve analysis was conducted at the end of the protocol to confirm amplification of a single product. β-actin was used as a housekeeping gene and gene expression levels were normalised to the dextrose-injected control group. Primer sequences were as follows: β-actin, Fwd 5′ GACCTCTATGCCAACACAGT 3′ and Rv 5′ AGTACTTGCGCTCAGGAGGA 3′; IL-6, Fwd 5′ ATGGATGCTACCAAACTGGA 3′ and Rv 5′ CCTCTTGGTTGAAGATATGA 3′; TNFα, Fwd 5′ CAGACCCTCACACTCAGATCATCT 3′ and Rv 5′ CCTCCACTTGGTGGTTTGCTA 3′; IL-1β, Fwd 5′ GGACAGAATATCAACCAACAAGTGATA 3′ and Rv 5′ GTGTGCCGTCTTTCATTACACAG 3′; IL-10, Fwd 5′ GGTTGCCAAGCCTTATCGGA 3′ and Rv 5′ ACCTGCTCCACTGCCTTGCT 3′; TGF-β, Fwd 5′ GACCAGCCGCCGCCGCAGG 3′ and Rv 5′ AGGGCTGTCTGGAGTCCTC 3′.

183

Confocal Raman mapping of tissue sections.

Mice were culled and spleens were immediately extracted and washed in 1X phosphate buffered saline (PBS) (Sigma, UK). One quarter of each extracted spleen was immediately transversely immersed into the optimal cutting temperature (OCT) media and then snap- frozen in isopropanol (Sigma, UK) pre-cooled to liquid nitrogen temperatures. The spleens were then cut into 5 µm sections via cryotomy (Leica, UK) and placed onto glass slides. The tissue sections were then fixed using 300 µL of ice cold methanol (Sigma, UK) for 10 min. The sections were subsequently gently washed in 1X PBS twice in order to remove the OCT. MilliQ water was then used to remove any salt crystals from the PBS. The sample was then re-doused with 300 µL of ice cold methanol and dried for 1 h at 37ºC to dehydrate the sample. Raman scanning was completed on regions of the spleen tissue including the red and white pulp regions. Raman scanning was completed using a polystyrene calibrated DXRi Raman Mapping system (Thermo Scientific, UK) using the following conditions: ʎ = 633 nm, 1 mW, objective 50x (numerical aperture: 0.75) pixel size = 3 µm. The instrument had an energy resolution of 2.5 cm-1. Dex 5% treated mice spleens were treated in the same way.

Immunofluorescence with Raman spectroscopy.

Mice were culled and the spleens were immediately extracted and washed in 1X PBS (Sigma, UK). One quarter of the spleen was immediately transversely immersed into the optimal cutting temperature (OCT) and snap-frozen in isopropanol (Sigma, UK) that was precooled to liquid nitrogen temperatures. The spleens were next sectioned into 5 µm sections via cryotomy, the sections were collected onto glass slides (VWR, UK). The tissue sections were afterwards fixed using 300 µL ice cold methanol (Sigma, UK) for 10 min. The methanol was removed and the fixed tissues were incubated in blocking buffer (1% bovine serum albumin (Sigma, UK), 0.1% Triton X100 in 1X PBS) for 1 h in a humidification chamber. Tissue sections were then washed in a washing buffer that consisted of 0.1% Triton X100 in 1X PBS, 3 times for 5 min. The washing buffer was removed and the tissue was incubated with the primary antibody, biotinylated Anti-SIGN Related 1 antibody [ER- TR9] AB51819 (Abcam, UK) diluted in blocking buffer to 3.3 µg/mL for 2 h at room temperature and pressure in a humidification chamber. The primary antibody solution was then removed and the tissue sections were gently washed with the washing buffer 3 times for 5 min each. The tissue sections were subsequently incubated with the revealing agent, Fluorescein Avidin DCS (Vector Laboratories, UK) which was diluted to 22 µg/mL with HEPES buffered saline (pH 8.2, 20°C), for 2 h in a humidification chamber. The tissue was then washed with washing buffer and excess liquid was removed using a cotton bud. The

184 tissue sections were next cover slipped with Prolong gold antifade including DAPI (Thermo Scientific, UK). The sample was left in darkness at room temperature for a 24 h period. Fluorescence microscopy was then completed and the regions where ERTR-9 staining was detected were marked on a customised grid. Secondary and negative controls were included in this experiment in order to exclude false positive readings. The coverslips were later gently removed by incubating the slides in 1X PBS overnight at room temperature. Following this treatment, the coverslips could be removed via gravity; care was taken not to damage the fixed tissue. The exposed tissues were gently washed with MilliQ water and left to dry at 37ºC for 1 - 2 h. Using the marked grids the same regions identified by fluorescence microscopy were analysed by a “point–and-shoot” polystyrene calibrated DXR Raman microscope (Thermo Scientific, UK) using a laser at λ = 633 nm, 0.4 mW, 50x (numerical aperture of 0.75) and an exposure time of 25 s, and averaged over 3 readings, the energy resolution of the instrument was 2.5 cm-1. This allowed the co-localisation of the Raman Signature of GO with the immunofluorescence signal of MZMs with bound ERTR-9 antibody. The spleens of at least 2 mice with 2 or more splenic sections were imaged per mouse.

Fluorescence activated cell sorting (FACS) with Raman spectroscopy.

A single cell suspension of splenocytes was obtained as described above. The isolated cells were then stained with the relevant antibodies mentioned below as per the manufacturer’s instructions. FACS was then performed on a BD Influx cell sorter (BD Biosciences, Oxford, UK). Fluorescently conjugated antibodies (Miltenyi Biotec Limited, UK) were excited and their emission was collected under the following conditions: CD45-VioBlue, violet laser, band pass 460/50; F4/80-v770-PE, yellow-green laser, 750LP; CD11c-FITC, blue laser, band pass 530/40; CD209-PE, yellow-green laser, band pass 585/29; CD169b, red laser, band pass 670/30. Cells were gated on an initial scatter gate to exclude debris. Lymphocytes were identified as CD45+/side scatter low and granulocytes were CD45+/side scatter high. F4/80+/CD11c- macrophages were gated from the granulocyte gate and CD11c+/F4/80- dendritic cells were gated from this population too. The CD209+ and CD169b+ cells were identified as being F4/80-/CD11c- and identified by plotting them against each other. Baseline voltages were set using unlabelled cells and positive populations were identified using full minus 1 (FMO) controls. Cells were filtered through a 50 µm mesh prior to being sorted through a 100 µm diameter nozzle at 20 PSI pressure, which generated a drop drive frequency of 38 KHz. Sort side streams were optimised to enable 5 separate populations to be collected simultaneously from each individual sample. Up to 20,000 cells were collected depending on the frequency of the cell sub population in the sample. The event rate was maintained at 2,000-3,000 events per second to minimise aberrant droplet formation and ensure a high degree of purity in the sorted cells. Cells were collected into 1.5 mL tubes

185 containing 1X PBS and then deposited onto glass slides via Cytospin apparatus using a centrifuge set to 350 g. The cells were fixed on to the glass slide with a drop of ice cold methanol. Once dry, the sample was analysed via Raman microscopy using a DXR Raman microscope (Thermo Scientific, UK) calibrated to polystyrene with an energy resolution of 2.5 cm-1. Measurements used a laser at λ = 633 nm, 0.4 mW, 50x (numerical aperture of 0.75), an exposure time of 25 s and averaged over 3 readings. The percentage of scanned cells in which GO was detected was determined at each time point, n = 3 mice x 20 cells. Cells isolated from Dex 5% treated controls were treated in the same way and scanned to ensure there was no cross contamination of GO between samples.

TUNEL staining.

Spleen sections were embedded in paraffin wax and sectioned to 5 µm thickness on a microtome (Leica, UK) and placed on glass slides. The sections then underwent deparafinisation via 2 immersion in xylene for 5 min each, followed by washing in 100% methanol for 5 min. The samples were subsequently rehydrated by passing them through a series of decreasing concentrations of alcohol concentrations (100%, 95%, 85%, 70%, 50%) for 3 minutes at each concentration. The sections were further washed in 0.85% NaCl for 5 min. The rehydrated sections were then fixed in 4% methanol-free paraformaldehyde for 15 min. The remaining paraformaldehyde was the removed by washing in PBS 1X (Sigma, UK) twice for 5 min each. The excess PBS was gently removed with filter paper being careful to not touch the sections. The glass slides with their fixed sections were then placed on a flat surface and 10 µL of a 20µg/ml proteinase K solution was placed on each tissue section. The slides were left with the proteinase K solution for an optimised time of 8 min. The proteinase K was removed by washing PBS for 5 minutes. The tissue sections were next fixed by immersion in 4% methanol-free paraformaldehyde for 5 min and then washed by immersion in PBS 1X (Sigma, UK) for 5 min. Then the tissue was covered with 100 µL of equilibrium buffer and left to equilibrate for 5 min. The equilibrium buffer was removed and 50 µL of the rTdT incubation buffer was added on to each slide and left for 60 min in a humidification chamber at 37ºC. The slides were next immersed in physiological saline (sodium chloride) 2X to terminate the reaction. The slides were then washed in PBS 1X (Sigma, UK) twice for 5 minutes and then washed twice in MilliQ water for another 5 min each. The slides were thereafter incubated in 4% Sudan black (in ethanol) for 10 minutes to eliminate red blood cell derived fluorescence, followed by washing 7 times in PBS 1X (Sigma, UK) for 3 minutes each and 3 washes in MilliQ water for 3 mins each. The slides were then cover slipped using with Prolong gold antifade including DAPI (Thermo Scientific, UK) and left at room temperature in darkness for 24 h. Images were collected using a 20x objective in a 3D Histech Panoramic 250 Flash slide scanner. Images were processed and

186 analysed using Pannoramic Viewer (http://www.3dhistech.com/) and Fiji/ImageJ software (version 1.5c; National Institutes of Health, Bethesda, MD). At least 3 mice were examined per condition. A total of 5 images of 300µm2 were analysed per mouse. The number of TUNEL positive cells were then blindly counted by 3 individuals and the counts of the 3 individuals were averaged for each image.

Raman spectroscopy on homogenised spleens.

For each time point analysed, a quarter of the extracted and washed individual murine spleens were mechanically homogenised on to a glass slide and left to dry for 30-60 min at room temperature and pressure. This was completed in order to assess the state of the overall GO content of the spleen, instead of cell specific content. Raman spectroscopy was then performed on the tissue homogenates using a polystyrene calibrated DXR Raman spectrometer (Thermo Scientific, UK) using a λ = 633 nm LASER a 50X objective lens (numerical aperture of 0.75) and a laser power of 0.4 mW with an exposure time of 25 s. The energy resolution was 2.5 cm-1. Raman scattering signals were acquired in at least 10 different regions per mouse with 3 biological replicas. Raman spectra were corrected for tissue auto-fluorescence using a 6th order polynomial baseline correction, and the ratio of intensity of the D and G peaks, I(D)/I(G), were calculated and averaged for n = 10 spectra for each of the 3 mice. We also scanned Dex 5% treated animals in the same way to ensure no cross contaminations as a quality check.

Histological marginal zone analysis.

For each time point in the degradation study, a quarter of the spleen was fixed with 4% paraformaldehyde and dehydrated through increasing concentrations of ethanol (50-100%) and embedded in paraffin wax. Sections of 5 µm thickness were cut and collected onto glass slides and then stained with H & E stain. The stained sections were then cover slipped with resinous mounting media and allowed to dry for 48 h at 37 ºC. Images were collected using a 20x objective and a 3D Histech Panoramic 250 Flash slide scanner. Images were processed and analysed to determine the average number of dark regions in the splenic marginal zone (n=5 marginal zones for each of the 3 mice) using Panoramic Viewer (http://www.3dhistech.com/) and Fiji/ImageJ software (version 1.5c; National Institutes of Health, Bethesda, MD).

TEM on splenic tissue sections.

For each time point analysed, a quarter of each of the extracted and washed murine spleens was fixed in paraformaldehyde (4% (w/v))/ glutaraldehyde (2.5% (w/v)) (Sigma, UK) for 24 h.

187

Following fixation, specimens were cut into 2-3 mm blocks using a sterile surgical blade. Each specimen was washed several times with deionised water, submitted to a second fixation with reduced osmium tetroxide (OsO4) (Agar Scientific, UK) for 90 min, then rinsed with deionised water and dehydrated through a series of graded ethanol (Fisher Scientific, UK) solution for 15 min at each concentration: 30%, 50%, 70%, 90% and 100%, dilutions were made in Milli Q water. This was followed by the incubation of the specimen in 100% acetone (Fisher Scientific, UK) for 15 min. Each specimen was infiltrated with increasing concentrations of polypropylene TAAB 812 hard resin (TAAB laboratories Ltd, UK) (25%, 50%, 75% and 100%, using acetone as the diluent) for 12 h at each grade. Specimens were next left open, in 100 % resin for a further 6 h at room temperature. Specimens were thereafter orientated in the embedding mould filled with fresh 100% resin, such that transversal sectioning of the spleen sample would be possible. The resin (with the orientated sample) was then cured at 60˚C for 48 h to allow resin polymerisation. TEM imaging was performed on ultrathin sections (approximately 60 nm thickness) obtained using an Ultracut E ultratome (Reichert-Jung, Austria) and a diamond knife (Diatome 45°, Leica, UK). Ultra- thin sections were collected onto coatless 200 mesh thin copper 3.05 mm grids (Electron Microscopy Services, USA) and observed under FEI Tecnai T-12 BioTWIN TEM (FEI, Eindhoven, NL) equipped with an Orius CCD SC100 camera (GATAN, UK) at 100 kV. The use of this microscope and grids allowed efficient observation, whereabouts and tracking of the GO within tissue. We also scanned Dex 5% treated animals in the same way to ensure no cross contaminations as a quality check.

Electron diffraction of GO in tissue sections.

Tissues were fixed and resin embedded as described above. TEM imaging was performed on ultrathin sections (approximately 60 nm thick) that were obtained using an Ultracut E ultratome (Reichert-Jung, Austria) and a diamond knife (Diatome 45°, Leica, UK). The ultrathin sections were collected onto Quantifoil S 7/2 copper grids (Electron Microscopy Services, USA). TEM and selected area electron diffraction (SAED) shown in Figures 6.5Aiv, 6.S13 and 6.S14 was performed on a FEI Talos 200X (FEI, Eindhoven, NL) operating at 80 kV and using electron dose rates ranging between 45 e-.A-1s-1 and 84 e-.A-1s- 1. TEM images and SAED patterns were acquired using a FEI Ceta CMOS camera on an area of 3.14 x 10-2 μm2 and 0.5 μm2 corresponding to the selected area apertures of 10 μm or 40 μm, respectively. An acquisition time of 1 s was used for the collection of TEM images and SAED patterns. However, for the acquisition of SAED patterns with a 10 μm selected area aperture, the acquisition time was set to of 8 s.

188

Statistical analysis.

All experimental data, are represented as mean ± standard error (SE) or mean ± standard deviation (SD) as stated in each section, with at least n=3-4. Statistical significance was determined by Kruskal–Wallis non-parametric test with Dunn’s post-hoc test. Or a Mann- Whitney U test for pairwise comparisons where appropriate. For RT-qPCR data Levene’s test for homogeneity of variances was performed, if it showed significance (p <0.05), then Welch ANOVA and Games-Howell’s post-hoc test was performed. SPSS software, version 22.0 was used to perform this analysis. In all experiments, probability values of <0.05 were considered significant.

5.7. Acknowledgements. This work was partially supported by the EPSRC NowNano DTC program and grants EP/K016946/1 and EP/M010619/1, the EU 7th RTD Framework Programme, Graphene Flagship project (FP7-ICT-2013-FET-F-604391) and DTRA, USA (grant HDTRA1-12-1-0013). The authors would like to acknowledge the staff in the Faculty of Biology, Medicine and Health EM Facility, particularly Dr Aleksandr Mironov and Ms Samantha Forbes for their expertise and assistance, and the Wellcome Trust for equipment grant support to the EM Facility. The University of Manchester Bioimaging Facility microscopes used in this study were purchased with grants from the BBSRC, Wellcome Trust and the University of Manchester Strategic Fund. The authors wish to acknowledge Mr R. Meadows from the Bio-imaging Facility. The authors are also thankful to Mr P. Walker, from the Histology Facility, University of Manchester for his expert advice and assistance in tissue histology. The authors thank Dr N. Hodson from the Bio-AFM Facility (Centre for Tissue Injury and Repair at the University of Manchester) for assistance and advice regarding the AFM instrumentation.

189

5.8. Supplementary information. Day 1 30 180 270 1 0.849 (ns) 0.237 (ns) 0.029 (*) 30 1.00 (ns) 1.00 (ns) 180 1.00 (ns) 270

This was completed following the application of the Kruskal-Wallis test and Dunn’s post hoc test. A p value of 0.05 was considered significant. All obtained p values are shown, where * indicates a p value <0.05, while “ns” signifies no significance.

Day Treatment groups P Value

1 GO Dex 5% 1.00 (ns) 30 GO Dex 5% 1.00 (ns) 180 GO Dex 5% 1.00 (ns) 270 GO Dex 5% 1.00 (ns)

Table 5.S2 Statistics table, showing p values, describing the differences between the apoptotic cell density (apoptotic cells / mm2) detected in spleen sections of mice treated with GO (7.5 mg /kg) treated or Dex 5% at different time points.

The p values shown were obtained following the application of the Man-Whitney U pairwise analysis, a p value of 0.05 was considered significant. ‘ns’ signifies no statistical significance. This table corresponds to Figure 5.4C.

190

Day 1 7 14 30 90 180 270 1 1.000 (ns) 1.000 (ns) 1.000 (ns) 0.801 (ns) 0.098 (ns) 0.026 (*) 7 1.000 (ns) 1.000 (ns) 1.000 (ns) 0.178 (ns) 0.052 (ns) 14 1.000 (ns) 1.000 (ns) 1.000 (ns) 0.627 (ns) 30 1.000 (ns) 1.000 (ns) 1.000 (ns) 90 1.000 (ns) 1.000 (ns) 180 1.000 (ns) 270

Table 5.S3 Statistics table (corresponding to Figure 5.5B), showing p values, illustrating the statistical differences in the number of dark areas per marginal zone between different time points following the administration of 7.5 mg/kg GO.

The p values shown derive from the application of the Kruskal-Wallis and Dunn’s post hoc test. A p value of 0.05 was considered significant. All obtained p values are shown, where * indicates a Value <0.05 while “ns” signifies no significance.

Day 1 7 14 30 90 180 270 1 1.000 (ns) 1.000 (ns) 0.007 (*) 0.016 (*) 0.018 (*) 0.001 (***) 7 1.000 (ns) 0.000 (***) 0.002 (***) 0.005 (**) 0.000 (***) 14 0.000 (***) 0.000 (***) 0.001 (***) 0.000 (***) 30 1.000 (ns) 1.000 (ns) 1.000 (ns) 90 1.000 (ns) 1.000 (ns) 180 1.000 (ns) 270

Table 5.S4 Statistics table, showing p values, describing the differences in the I(D)/ I(G) ratios between different time points following the intravenous administration of 7.5 mg/kg of GO.

These results were obtained following the application of the Kruskal-Wallis and Dunns post hoc test. A p value of 0.05 was considered significant. All obtained p values are shown, where *, ** and *** denotes a p value < 0.05, < 0.01 and < 0.005, respectively. “ns” signifies no significance. This table corresponds to the data presented in Figure 5.5C.

191

A B 100

80 27%

60 13%

40 % Weight %

Intensity(a.u.) 20

0 4000 3000 2000 1000 200 400 600 Wavenumber (cm-1) Temperature (oC)

50 C1s Scan

) ) C ii) 3

i) 4 O1s C-C/ C=C 60 40 C-O C=O O=C-O

40 C1s 30 

CPS x 10 x CPS CPS x 10 x CPS

( Background ( 20 Envelope 20

10 Intensity Intensity Intensity 0 0 1200 900 600 300 0 292 290 288 286 284 282 280 Binding Energy (eV) Binding Energy (eV) [GO] g/mL 0.6 R2 = 0.99 80 R2 = 0.99 [GO] g/mL 7.5 100 y = 0.029x E y = 0.36x 75 D 0.6 10.0 100 0.4 60 12.5 75 125 15.0 40 150 0.4 0.2 17.5 175 10 15 20 20.0 50 100 150 200 [GO] (g/mL) [GO] (g/mL) 200 0.2 25

0.0 0 Absorption Intensity Absorption 200 300 400 500 600 Intensity Fluorescence 600 700 800 900 Wavelength (nm) Wavelength (nm)

Figure 5.S1 Physicochemical characterisation of starting GO material.

(A) FT-IR absorbance spectra (B) TGA curve with % weight loss at the 2 main weight loss steps inset. (C) XPS analyses showing i) XPS survey spectra in and ii) the C1s XPS spectra. XPS was performed at NEXUS and Dr Neus Lozano analysed the data. (D) UV-vis absorption spectra with the concentration dependent calibration curve (λ max (abs) = 230 nm). (E) Fluorescence spectra with the concentration dependent calibration curve (λ max (em) = 595 nm).

192

Figure 5.S2 SAED aperture micrographs of the region of interest corresponding to the electron diffractogram in Figure 5.1. The scale bar represents 200 nm.

Dex 5% GO 2.5 mg/kg GO 5 mg/kg GO 10mg/kg 24 h 24

1 month 1

Figure 5.S3 The gross appearance of the spleens of C57BL/6 mice 24 h and 1 month after injection with GO at different concentrations (2.5, 5 and 10 mg/mL), compared to those of control mice (mice injected with Dex 5%).

193

24 h 1 month

Dex

GO 2.5 mg/kg 2.5 GO

GO 5 mg/kg 5 GO

GO 10 mg/kg 10 GO LPS 5 mg/kg 5 LPS

Figure 5.S4 Murine spleen sections that were stained with the TUNEL stain at 24 and 1 month post GO administration with at different concentrations (2.5, 5, 10 mg/kg) or Dex 5%.

Dashed white lines indicate the boundaries of the marginal zone which separates red and white pulp where present. Scale bars represent 50 μm. Yein Nam performed this experiment and the data was analysed by Dr Dhifaf Jasim who also contributed the figure.

194

500 24h 1 month 400 LPS at 24h ***

300

) 2 *** mm 200 ***

100 * Apoptotic / (cellscelldensity Apoptotic 0 Dex 5% LPS 5 mg/kg GO 2.5 GO 5 mg/kg GO 10 mg/kg mg/kg

Figure 5.S5 Apoptotic cells density as per TUNEL assay at 24 and 1 month post-injection with GO at different concentrations (2.5, 5, 10 mg/kg) or Dex 5%.

Statistical significance were assessed via the Kruskall – Wallis test with Dunns Post hoc test (p < 0.05 *, p < 0.005 *** vs. age matched negative control). Mice were n = 3-4 mice per group. Dr Dhifaf Jasim conducted the experiments and contributed this Figure.

195

24 h 1 month

20.7% 16.5%

CD4 CD4

CD4 CD4 Dex

17.3% 13.4%

CD8 CD8

22.4% 19.3%

CD4 CD4 CD4 CD4

15.9% 12.7% mg/ml 2.5 GO

CD8 CD8

27.1% 21.2%

CD4 CD4

CD4 CD4 CD4 CD4

16.9% 12.9% mg/ml 5 GO

CD8 CD8

22.7% 15.1%

CD4 CD4 CD4 CD4

22.2% 13.4% GO 10 mg/ml 10 GO

CD8 CD8

13.3%

CD4 CD4 LPS 5 mg/kg 5 LPS 11.8%

CD8 CD8

Figure 5.S6 Representative flow cytometry plots outlining the CD4 and CD8 cell counts of mice 24 h and 1 month after administration of GO at different concentrations (2.5, 5 and 10 mg/mL), compared to those obtained for control mice, Dex 5% ( negative control and LPS-injected mice at 24 h (positive control).

Dr Dhifaf Jasim conducted the experiments and contributed this Figure.

196

20 GO Treated

Dex 5% Treated Weight of Mouse (g) Mouse of Weight

10 0 100 200 300 Time (Days)

Figure 5.S7 Growth curve of mice injected with GO (7.5 mg/kg) or Dex 5% (control) over 9 months.

Data are represented by mean ± standard error (SE). Each time point is an average of the weights of n = 5-10 mice.

197

GO 7.5mg/kg Dex 5%

Day 1 Day

Day 30 Day Day 270 Day

Figure 5.S8 Murine spleen sections stained with TUNEL stain at different time points (Day 1, 30 and 270) after GO intravenous administration (7.5 mg/kg).

Optical images of stained spleen sections that had been extracted from mice treated with GO 150µg or Dex 5% are provided. Dashed white lines indicate the boundaries of the marginal zone which separates the red and white pulp where present. Scale bars represent 50 μm.

198

60 * Day 1 Day 30 Day 180 40 Day 270

0 screened positive for GO Raman signature signature Raman GO for positive screened % of cells of individual cell populations that that populations cell individual of cells of % Red Pup Lymphocytes Dendritic CellsMarginal Zone Metallophillic Macrophages Macrophages Macrophages

Figure 5.S9 Cellular distribution of splenic GO at different time points following GO administration (7.5 mg/kg).

The percentage of cells of individual spleen cell populations that were found positive for GO over time (Days 1, 30,180, 270). Data was acquired via FACS followed by Raman spectroscopy of the isolated cells. Statistical significance was assessed using a Kruskal-Wallis test followed by Dunns post hoc test (p < 0.05), 20 cells were counted per mouse and there were n = 3 mice per group.

199

600 Day 1

400

200

0 1000 1500 2000 Raman Shift (cm-1) 600 Day 30 400

200

0 1000 1500 2000 -1

.) Raman Shift (cm ) a.u 600 Day 90

400

200

0 1000 1500 2000 Raman Shift (cm-1)

Raman Intensity ( Intensity Raman 600 Day 180

400

200

0 1000 1500 2000 Raman Shift (cm-1)

600 Day 270

400

200

0 1000 1500 2000 Raman Shift (cm-1)

Figure 5.S10. The evolution of the average Raman spectra of splenic captured GO over 9 months.

Raman spectra of GO in physically homogenised tissue at different time points, Day 1, 30, 90, 180 and 270, n =10 spectra were collected for each of the 3 spleens extracted from 3 treated animals per time point.

200

GO 7.5mg/kg Dex 5%

Day 1 Day

Day 30 Day

Day 90 Day

Day 180 Day Day 270 Day

Figure 5.S11 Representative H & E stained splenic sections of mice at different time points following GO administration (7.5 mg/kg).

Spleens of mice treated with GO treated mice are shown on the right and those of mice treated with Dex 5% treated, are shown on the left. Scale bars represent 50 µm. Inset images show the presence of GO within the vicinity of cells of the marginal zone, scale bars represent 10 µm.

201

GO 7.5mg/kg Dex 5%

Day 1 Day

Day 30 Day

Day 90 Day

Day 180 Day

270 270 Day Day

Figure 5.S12 TEM micrographs of GO sequestered within the vesicular compartments of murine marginal zone splenocytes at different time points following administration of 7.5 mg/kg GO.

Dex 5% treated controls are also shown (right), Scale bars represent 1 µm. Insets show magnifications of vesicle entrapped GO material in GO treated mice spleens or equivalent regions within Dex 5% treated mice spleens at the respective time points. Scale bars represent 500 nm.

202

Corresponding Diffractograms Profile micrographs

Day 1 Day 2

0 4 6 8 10 Reciprocal Distance (1/nm) 6

2 Day 30 Day

0 4 6 8 10

Reciprocal Distance (1/nm) ) 4 6

2 Day 90 Day

0 4 6 8 10 Reciprocal Distance (1/nm)

6 Intensity (counts x10 (counts Intensity

2 Day 180 Day 0 4 6 8 10 Reciprocal Distance (1/nm)

2 Day 270 Day 0 4 6 8 10 Reciprocal Distance (1/nm)

Figure 5.S13 Representative electron diffractograms of intracellularly entrapped GO material within the spleens of mice at different time points following administration of GO (7.5 mg/kg), with the corresponding intensity line profiles.

The corresponding micrographs at different time points are provided, scale bars represent 200nm. Dotted circles indicate regions where electron diffraction patterns were acquired. Dr Eric Prestat obtained the electron diffraction patterns shown here.

203

Low magnification High magnification

Figure 5.S14 Holes seen within intracellularly entrapped GO sheets, which are indicative of an early transitional state of GO as its biodegradation proceeds.

Such structures were predominantly seen at 30 days post GO administration (7.5 mg/kg). Low magnification, scale bars represent 500 nm, and high magnification, scale bars represent 50 nm, are provided. The electron diffraction patterns of the structures are also given as insets where possible, where scale bars represent 10 1/nm. Dr Eric Prestat obtained the electron diffraction patterns shown here.

204

A 5 nm 100 800

80 600 60 400 40 200

20 650) = (n Count Count (n = (n Count 650) 0 0 0 1 2 3 4 0 200 400 600 Thickness (nm) Lateral Dimension (nm) 1 μm 0 nm

B C D

I(D)/ I(G) = 1.46 + 0.15

500 nm Intensity Raman (a.u.) 1000 2000 3000 10 1/nm Raman Shift (cm-1)

Figure 5.S15 Physicochemical characterisation of GO material after 9 months incubation in water at 37 ̊C in darkness.

Structural characterisations were completed using (A) AFM, where images were scrutinised for lateral dimensions and thickness after counting 650 individual GO sheets. (B) TEM, a representative TEM micrograph is provided (C) Raman spectroscopy with the I(D)/I(G) ratio provided in the inset. (D) Electron diffraction. The corresponding SAED aperture micrograph is shown in Figure 5.S16. Dr Eric Prestat obtained the electron diffraction patterns shown here.

205

Figure 5.S16 SAED aperture micrographs of the region of interest corresponding to the electron diffractogram in Figure 5.1.

Scale bar represents 50 nm.

206

5.9. References.

1 Novoselov, K. S. et al. A roadmap for graphene. Nature 490, 192-200 (2012). 2 Kostarelos, K. & Novoselov, K. S. Graphene devices for life. Nat. Nanotechnol. 9, 744-745 (2014). 3 Sun, X. et al. Nano-graphene oxide for cellular imaging and drug delivery. Nano. Res. 1, 203-212 (2008). 4 Kostarelos, K. & Novoselov, K. S. Materials science. Exploring the interface of graphene and biology. Science 344, 261-263 (2014). 5 Girish, C. M., Sasidharan, A., Gowd, G. S., Nair, S. & Koyakutty, M. Confocal raman imaging study showing macrophage mediated biodegradation of graphene in vivo. Adv. Healthc. Mater. 2, 1489-1500 (2013). 6 Yang, K. et al. In vivo pharmacokinetics, long-term biodistribution,and toxicology of pegylated graphene in mice. ACS Nano 5, 516-522 (2011). 7 Bussy, C., Jasim, D., Lozano, N., Terry, D. & Kostarelos, K. The current graphene safety landscape - a literature mining exercise. Nanoscale 7, 6432-6435 (2015). 8 Jasim, D. A., Menard-Moyon, C., Begin, D., Bianco, A. & Kostarelos, K. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. Chem. Sci. 6, 3952-3964 (2015). 9 Aichele, P. et al. Macrophages of the splenic marginal zone are essential for trapping of blood-borne particulate antigen but dispensable for induction of specific T cell responses. J. Immunol. 171, 1148-1155 (2003). 10 Liu, D., Mori, A. & Huang, L. Large liposomes containing ganglioside GM1 accumulate effectively in spleen. Biochim. Biophys. Acta. 1066, 159-165 (1991). 11 Mebius, R. E. & Kraal, G. Structure and function of the spleen. Nat. Rev. Immunol. 5, 606-616 (2005). 12 Zhu, J. & Paul, W. E. CD4 T cells: fates, functions, and faults. Blood 112, 1557-1569 (2008). 13 Harty, J. T., Tvinnereim, A. R. & White, D. W. CD8+ T cell effector mechanisms in resistance to infection. Annu. Rev. Immunol. 18, 275-308 (2000). 14 Wick, P. et al. Classification framework for graphene-based materials. Angew. Chem. Int. Ed. 53, 7714-7718 (2014). 15 Kurapati, R. et al. Dispersibility-dependent biodegradation of graphene oxide by myeloperoxidase. Small 11, 3985-3994 (2015). 16 Kotchey, G. P. et al. The enzymatic oxidation of graphene oxide. ACS Nano 5, 2098- 2108 (2011).

207

17 Bai, H. et al. Insight into the mechanism of graphene oxide degradation via the photo-fenton reaction. J. Phys. Chem. C. Nanomater. Interfaces 118, 10519-10529 (2014). 18 Zhou, X. et al. Photo-fenton reaction of graphene oxide: a new strategy to prepare graphene quantum dots for DNA cleavage. ACS Nano 6, 6592-6599 (2012). 19 Xing, W., Lalwani, G., Rusakova, I. & Sitharaman, B. Degradation of graphene by hydrogen peroxide. Part. Part. Syst. Charact. 31, 745-750 (2014). 20 Lalwani, G., Xing, W. & Sitharaman, B. Enzymatic degradation of oxidized and reduced graphene nanoribbons by lignin peroxidase. J. Mater. Chem. B 2, 6354- 6362 (2014). 21 Ding, Y., Tian, R., Yang, Z., Chen, J. & Lu, N. NADPH oxidase-dependent degradation of single-walled carbon nanotubes in macrophages. J. Mater. Sci. Mater. Med. 28, 7 (2017). 22 Elgrabli, D. et al. Intracellular degradation of functionalized carbon nanotube/iron oxide hybrids is modulated by iron via Nrf2 pathway. Sci. Rep. 7, 40997 (2017). 23 Lu, N., Li, J., Tian, R. & Peng, Y.-Y. Binding of human serum albumin to single- walled carbon nanotubes activated neutrophils to increase production of hypochlorous acid, the oxidant capable of degrading nanotubes. Chem. Res. Toxicol. 27, 1070-1077 (2014). 24 Kendall, M. & Lynch, I. Long-term monitoring for nanomedicine implants and drugs. Nat. Nanotechnol. 11, 206-210 (2016). 25 Russell, E. S., Neufeld, E. F. & Higgins, C. T. Comparison of normal blood picture of young adults from 18 inbred strains of mice. Proc. Soc. Exp. Biol. Med. 78, 761-766 (1951). 26 Jasim, D. A. et al. The effects of extensive glomerular filtration of thin graphene oxide sheets on kidney physiology. ACS Nano 10, 10753-10767 (2016). 27 Ferrari, A. C. & Robertson, J. Interpretation of Raman spectra of disordered and amorphous carbon. Phys. Rev. B 61, 14095-14107 (2000). 28 Georgakilas, V. et al. Functionalization of graphene: covalent and non-covalent approaches, derivatives and applications. Chem. Rev. 112, 6156-6214 (2012). 29 Vacchi, I. A., Spinato, C., Raya, J., Bianco, A. & Menard-Moyon, C. Chemical reactivity of graphene oxide towards amines elucidated by solid-state NMR. Nanoscale 8, 13714-13721 (2016). 30 Bianco, A. et al. All in the graphene family – A recommended nomenclature for two- dimensional carbon materials. Carbon 65, 1-6 (2013).

208

31 Illum, L. et al. Blood clearance and organ deposition of intravenously administered colloidal particles. The effects of particle size, nature and shape. Int. J. Pharm. 12, 135-146 (1982). 32 Bussy, C., Ali-Boucetta, H. & Kostarelos, K. Safety considerations for graphene: lessons learnt from carbon nanotubes. Acc. Chem. Res. 46, 692-701 (2013). 33 Raghav, N., Chakraborty, S. & Maiti, P. K. Molecular mechanism of water permeation in a helium impermeable graphene and graphene oxide membrane. Phys. Chem. Chem. Phys. 17, 20557-20562 (2015). 34 Chen, S. et al. Mass spectrometry imaging reveals the sub-organ distribution of carbon nanomaterials. Nat. Nanotechnol. 10, 176-182 (2015). 35 Sasidharan, A. et al. Comparative in vivo toxicity, organ biodistribution and immune response of pristine, carboxylated and pegylated few-layer graphene sheets in swiss albino mice: a three month study. Carbon 95, 511-524 (2015). 36 Schnitzer, B., Rucknagel, D. L., Spencer, H. H. & Aikawa, M. Erythrocytes: pits and vacuoles as seen with transmission and scanning electron microscopy. Science 173, 251-252 (1971). 37 Holroyde, C. P. & Gardner, F. H. Acquisition of autophagic vacuoles by human erythrocytes. Physiological role of the spleen. Blood 36, 566-575 (1970). 38 Dameshek, W. Hypersplenism. Bull. N. Y. Acad. Med. 31, 113-136 (1955). 39 William, B. M. & Corazza, G. R. Hyposplenism: a comprehensive review. Part I: basic concepts and causes. Hematology 12, 1-13 (2007). 40 Doll, D. C., List, A. F. & Yarbro, J. W. Functional hyposplenism. South. Med. J. 80, 999-1006 (1987). 41 Gottlieb, Y. et al. Physiologically aged red blood cells undergo erythrophagocytosis in vivo but not in vitro. Haematologica 97, 994-1002 (2012). 42 Knutson, M. & Wessling-Resnick, M. Iron metabolism in the reticuloendothelial system. Crit. Rev. Biochem. Mol. Biol. 38, 61-88 (2003). 43 Kohyama, M. et al. Role for spi-c in the development of red pulp macrophages and splenic iron homeostasis. Nature 457, 318-321, (2009). 44 Franken, L. et al. Splenic red pulp macrophages are intrinsically superparamagnetic and contaminate magnetic cell isolates. Sci. Rep. 5, 12940 (2015). 45 Qu, G. et al. The ex vivo and in vivo biological performances of graphene oxide and the impact of surfactant on graphene oxide's biocompatibility. J. Environ. Sci. 25, 873-881 (2013). 46 Wu, H. et al. Hyaluronic acid conjugated graphene oxide for targeted drug delivery. Carbon 69, 379-389 (2014).

209

47 Kanakia, S. et al. Dose ranging, expanded acute toxicity and safety pharmacology studies for intravenously administered functionalized graphene nanoparticle formulations. Biomaterials 35, 7022-7031 (2014). 48 Nurunnabi, M. et al. In vivo biodistribution and toxicology of carboxylated graphene quantum dots. ACS Nano 7, 6858-6867 (2013). 49 Chong, Y. et al. The in vitro and in vivo toxicity of graphene quantum dots. Biomaterials 35, 5041-5048 (2014). 50 Zhang, D. et al. The short- and long-term effects of orally administered high-dose reduced graphene oxide nanosheets on mouse behaviors. Biomaterials 68, 100-113 (2015). 51 Nolan, Y., Vereker, E., Lynch, A. M. & Lynch, M. A. Evidence that lipopolysaccharide- induced cell death is mediated by accumulation of reactive oxygen species and activation of p38 in rat cortex and hippocampus. Exp. Neurol. 184, 794-804 (2003).

52 de Porto, A. P. N. A. et al. Assessment of splenic function. Eur. J. Clin. Microbiol. Infect. Dis. 29, 1465-1473 (2010). 53 Langeveld, M., Gamadia, L. E. & ten Berge, I. J. T-lymphocyte subset distribution in human spleen. Eur. J. Clin. Invest. 36, 250-256 (2006). 54 Wang, X., Podila, R., Shannahan, J. H., Rao, A. M. & Brown, J. M. Intravenously delivered graphene nanosheets and multiwalled carbon nanotubes induce site- specific Th2 inflammatory responses via the IL-33/ST2 axis. Int. J. Nanomedicine 8 (2013). 55 Orecchioni, M. et al. Molecular and genomic impact of large and small lateral dimension graphene oxide sheets on human immune cells from healthy donors. Adv. Healthcare Mater. 5, 276-287 (2016). 56 Sydlik, S. A., Jhunjhunwala, S., Webber, M. J., Anderson, D. G. & Langer, R. In vivo compatibility of graphene oxide with differing oxidation states. ACS Nano 9, 3866- 3874 (2015). 57 Assoian, R. K. et al. Expression and secretion of type beta transforming growth factor by activated human macrophages. Proc. Natl. Acad. Sci. U. S. A. 84, 6020-6024 (1987). 58 Said, E. A. et al. Programmed death-1-induced interleukin-10 production by monocytes impairs CD4+ T cell activation during HIV infection. Nat. Med. 16, 452- 459 (2010). 59 Lopez-Castejon, G. & Brough, D. Understanding the mechanism of IL-1beta secretion. Cytokine Growth Factor Rev. 22, 189-195 (2011).

210

60 Malstrom, C. & James, S. Inhibition of murine splenic and mucosal lymphocyte function by enteric bacterial products. Infect. Immun. 66, 3120-3127 (1998). 61 Ertel, W. et al. Downregulation of proinflammatory cytokine release in whole blood from septic patients. Blood 85, 1341-1347 (1995). 62 Trinschek, B. et al. Kinetics of IL-6 production defines T effector cell responsiveness to regulatory T cells in multiple sclerosis. PloS One 8, e77634 (2013). 63 Burkholder, T., Foltz, C., Karlsson, E., Linton, C. G. & Smith, J. M. Health evaluation of experimental laboratory mice. Curr. Protoc. Mouse Biol. 2, 145-165 (2012). 64 Wen, K. P. et al. Accumulation and toxicity of intravenously-injected functionalized graphene oxide in mice. J. Appl. Toxicol. 35, 1211-1218 (2015). 65 Yang, K. et al. In vivo biodistribution and toxicology of functionalized nano-graphene oxide in mice after oral and intraperitoneal administration. Biomaterials 34, 2787- 2795 (2013). 66 Li, B. et al. Biodistribution and pulmonary toxicity of intratracheally instilled graphene oxide in mice. NPG Asia Mater. 5, e44 (2013). 67 Suttie, A. W. Histopathology of the spleen. Toxicol. Pathol. 34, 466-503 (2006). 68 Masuda, T., Satodate, R., Tsuruga, K. & Kasai, T. Quantitative assessment of a change of hemosiderin deposition with age in splenic compartments of rats. Tohoku J Exp Med. 170, 169-179 (1993). 69 Seiler, P. et al. Crucial role of marginal zone macrophages and marginal zone metallophils in the clearance of lymphocytic choriomeningitis virus infection. Eur. J. Immunol. 27, 2626-2633 (1997). 70 Gherman, C. et al. Pharmacokinetics evaluation of carbon nanotubes using FTIR analysis and histological analysis. J. Nanosci. Nanotechnol. 15, 2865-2869 (2015). 71 Birjandi, S. Z., Ippolito, J. A., Ramadorai, A. K. & Witte, P. L. Alterations in marginal zone macrophages and marginal zone B cells in old Mice. J. Immunol. 186, 3441- 3451 (2011). 72 Schultz, Z. D., Stranick, S. J. & Levin, I. W. Tip Enhanced raman spectroscopy and imaging: an apical illumination geometry. Appl. Spectrosc. 62, 1173-1179 (2008). 73 Sharron, M. et al. Platelets induce apoptosis during sepsis in a contact-dependent manner that is inhibited by GPIIb/IIIa blockade. PloS One 7, e41549 (2012). 74 Figgett, W. A. et al. The TACI receptor regulates T-cell-independent marginal zone B cell responses through innate activation-induced cell death. Immunity 39, 573-583 (2013). 75 Thomsen, C. & Reich, S. Double resonant Raman scattering in graphite. Phys. Rev. Lett. 85, 5214-5217 (2000).

211

76 Ali-Boucetta, H. et al. Graphene oxide: purified graphene oxide dispersions lack In vitro cytotoxicity and in vivo pathogenicity. Adv. Healthc. Mater. 2, 512-512 (2013). 77 Jasim, D.A., Neus, L. & Kostas, K. Synthesis of few-layered, high-purity graphene oxide sheets from different graphite sources for biology. 2D Mater. 3, 014006 (2016). 78 Lucchese, M. M. et al. Quantifying ion-induced defects and raman relaxation length in graphene. Carbon 48, 1592-1597 (2010). 79 Smith, B. W. & Luzzi, D. E. Electron irradiation effects in single wall carbon nanotubes. J. Appl. Phys. 90, 3509-3515 (2001). 80 Elgrabli, D. et al. Carbon nanotube degradation in macrophages: live nanoscale monitoring and understanding of biological pathway. ACS Nano 9, 10113-10124 (2015). 81 Kagan, V. E. et al. Carbon nanotubes degraded by neutrophil myeloperoxidase induce less pulmonary inflammation. Nat. Nanotechnol. 5, 354-359 (2010). 82 Osmond-McLeod, M. J. et al. Durability and inflammogenic impact of carbon nanotubes compared with asbestos fibres. Part Fibre Toxicol. 8, 15 (2011). 83 Lammers, T., Kiessling, F., Hennink, W. E. & Storm, G. Drug targeting to tumors: principles, pitfalls and (pre-) clinical progress. J. Control. Release 161, 175-187 (2012). 84 Lipson, R. L., Bayrd, E. D. & Watkins, C. H. The postsplenectomy blood picture. Am. J. Clin. Pathol. 32, 526-532 (1959).

212

Chapter 6

6. Nose to Brain Translocation and Biodegradation of Graphene Oxide Nanosheets

Leon Newman1,2, Artur Filipe Rodrigues1,2, Dhifaf Jasim1,2, Isabella Anna Vacchi3, Alberto Bianco3, Cyrill Bussy*1,2 & Kostas Kostarelos1,2*

Affiliations: 1 Nanomedicine Laboratory, School of Health Sciences, Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, M13 9PT, UK. 2 National Graphene Institute, The University of Manchester, Manchester, M13 9PL, UK. 3 CNRS, Institut de Biologie Moléculaire et Cellulaire, Laboratoire d'Immunopathologie et Chimie Thérapeutique, F-67000 Strasbourg, France.

6.1. Statement.

Leon Newman, Dr Cyrill Bussy and Prof Kostas Kostarelos planned the study, designed the experiments and critically reviewed the manuscript. Leon Newman wrote the manuscript. Leon Newman performed the radiolabelling experiments and GO f3 materials characterisation (before and after functionalisation). He also performed the animal experiments and analytical work and data analysis regarding the investigation of the brain biodistribution and in vivo biodegradation of graphene oxide. Filipe Rodrigues synthesised the graphene oxide and critically reviewed the manuscript. He also helped with the animal experiments. Dhifaf Jasim operated and generated the SPECT/CT data; she also critically reviewed the manuscript. Isabella Anna and Prof Alberto Bianco functionalised the GO material with DOTA chelator groups. This manuscript has been written according to the specifications of the Chemical Science journal

______

* Correspondence to either: [email protected]

[email protected]

213

6.2. Abstract.

The nose-to-brain pathway is a distinctive route by which nanomaterials can gain access to the central nervous system, however the extent to which this happens for different nanomaterials and the fate thereof is still being explored. Graphene oxide (GO) is among those materials that have yet to be explored in this context. The need to understand the interactions of GO with biological systems and for human exposure is critical in light of the intensive research, which has revealed some promising applications for GO, which is gradually helping to narrow the gap between the laboratory bench-top GO research and GO marketed products. Using a range of analytical techniques, here we characterise the extent that precision engineered, extensively characterised GO sheets of different controlled lateral dimensions could translocate from the nose to the brain following their intranasal instillation. We then explore the tissue location and in vivo biodegradability of the translocated materials. Our results, from mass spectrometry and confocal Raman based analysis, indicate that GO sheets undergo nose-to-brain translocation in a limited, though size-dependent manner. The smallest sheet size category of GO (us-GO, 30 – 500 nm) was shown to gain greatest access to the brain in terms of quantity and brain coverage. Based on confocal Raman mapping and immunofluorescence combinations, we identify that once in the brain, us-GO sheets reside in close association to microglia, the brain’s resident macrophages. The quantity of us-GO within the brain decreases over time, and this is accompanied by biodegradative changes, as characterised via Raman spectroscopy. This study adds to the growing and necessary awareness of the biological interactions and fate of graphene-based materials with biological systems.

6.3. Introduction.

In recent years, the family of sp2 hybridised carbon nanomaterials (CNMs) has moved towards the forefront of nanotechnology for the development of new consumer products.1 Graphene is a two dimensional (2D) sheet of sp2 hybridised carbon atoms, which can be considered as the archetypal member of the CNM family. This material has received considerable attention as a result of the appreciation of its unique and versatile properties.2, 3 Since the discovery and characterisation of graphene4, many graphene-based materials (GBMs), including the pristine graphene sheets, single and few layer graphene oxide (GO) sheets, and graphene quantum dots have been described and investigated.5 GO, an oxidised derivative of graphene containing a myriad of oxygen functionalities,6 has been intensively studied for a variety of applications. These include chemically resistant coatings

214 with anti-corrosive properties7, membranes for the tuneable sieving of ions for water desalination8, energy storage9 as well as biomedical applications such as drug delivery, tissue engineering and diagnosis.10, 11 With the increased research and understanding of GBMs, the gap between the laboratory bench-top and consumer products is closing, making the need to understand the interactions of these unique materials with biological systems ever more critical.12

The nasal route represents a major means by which particles and nanomaterials can gain access to exposed individuals.13 As per the ICRP model of fractional depositions of inhaled particles14, the aerodynamic diameter of particles can be a major influence in determining the location of where materials deposit in the pulmonary tract. The smallest of particles, including some nanomaterials, deposit predominantly in the nasopharyngeal and laryngeal regions. Considering the unique anatomy of the olfactory region of the nose, which connects directly and indirectly with the brain15, this deposition may result in the nose to-brain translocation of certain nanomaterials. In support of this, various reports including epidemiologic studies, clinical trials as well as animal experiments concur that inhaled nanosized particles can be found in organs besides the respiratory tract, including the brain, following inhalation.16-18

The translocation and vulnerability of the central nervous system (CNS) to engineered nanoparticles was elegantly demonstrated in a landmark study by De Lorenzo, where gold particles were imaged in transit from the nasal mucosa into the brain using electron microscopy.19 Nose-to-brain transport is nowadays an established phenomenon. It has been reported by many research groups for a variety of nanoparticles, albeit in minute quantities, including silver nanoparticles20, iron (II) oxide nanoparticles21, ultrafine carbon particles22, titanium dioxide nanoparticles23, manganese oxide nanoparticles24, and exosomes.25 Recently and importantly, exogenous combustion derived iron oxide nanoparticles have been detected in the brains of deceased humans previously living in several distinct geographic locations.26 The observation of these nanomaterials in the brain may have been the result of nose-to-brain translocation following the inhalation of polluted air, as speculated by the authors.26

Several modes of transport have been considered by which nanomaterials could enter the brain from the nasal cavities, including transport through the axons of olfactory neurones (olfactory neural pathway)19 and trigeminal neurones (trigeminal pathway)27, 28, or paracellular transport between the axons of neurones.29 Other more rapid pathways include paracellular or transcellular transport through the olfactory sustentacular epithelial cells.15, 30, 31 In addition

215 to this, nanoparticles could undergo absorption into the systemic circulation followed by permeation of the blood brain barrier (BBB) to access the brain (systemic pathway).15 Although possible, the latter pathway remains the most unlikely due to various defences of the healthy BBB, such as efflux pumps and narrow tight junctions32, as well as the mononuclear phagocytic system. These prevent most exogenous materials from crossing from the systemic circulation into the brain, unless there is specific engineering of the materials, such as with antibodies or other targeting ligands that allow them to circumvent the BBB.33 Following translocation by one of these routes, the fate of nanomaterials within the brain remains an area of research that has only begun to be explored.18

While studies examining the potential of GBMs to cause harm to the pulmonary system exist34, 35, studies to explore the translocation from the nose to the brain are absent from the literature. This presents an important void that is imperative to fill, considering the recent surge in graphene research and applications that increases the potential for public and worker exposure. This is particularly true with regards to the use of GO in coatings that are likely to be applied to surfaces and will be subject to aging and possible damage.7

In order to address this knowledge gap, we designed a study where animals were intranasally exposed to aqueously dispersed GO sheets of different controlled lateral dimensions (large (l), small (s), and ultra-small (us) GO). Our specific objectives were to understand: i) the influence of lateral dimensions of GO sheets on the extent of nose-to-brain translocation, and ii) the biodegradability of the translocated material over a one month period. It was our hypothesis that the lateral dimensions of GO sheets will primarily govern the extent of translocation from the nasal cavities to the brain. Due to the highly oxidised and defected nature of the GO materials, we postulated that translocated GO sheets would subsequently be biodegraded by resident phagocytic immune cells, i.e. microglial cells or perivascular macrophages. Biodegradation of CNMs is a recent concept, however studies conducted by us and others have shown, both in vivo 36-39 [Chapter 5] and in vitro 40-42 that CNMs can undergo biologically mediated degradation under specific conditions.

Our results indicate that GO sheets undergo nose-to-brain translocation in a size-dependent manner, with us-GO sheets being the most prominently translocated material. The majority of the translocated materials were detected within the olfactory bulbs and associated predominantly with phagocytic microglial cells. The total quantity of material that translocated however was minuscule. Moreover, over the course of a month, the translocated GO underwent a slow but continuous biodegradative process as evidenced by Raman spectroscopy. Even after multiple intranasal instillations of 30 μg of GO (5 doses over 5 days, 1 dose/day), a similar pattern of distribution (predominantly in the olfactory bulbs and

216 to a lesser extent in more distal structures) and fate of materials (continuous degradation) was observed. This lack of difference between single and multiple exposures was attributed to the minor increase in the amount of materials translocated into the brain after repeated exposure.

217

6.4. Results.

Physicochemical characterisation of GO materials.

Graphene oxide (GO) of differing controlled lateral dimensions (l-GO, s-GO and us-GO) were prepared by a modified Hummers’ method under pyrogen-free conditions, as previously described and therefore suitable for biological investigations43, 44 [Rodrigues A. F., Newman L., Lozano N, Bussy C & Kostarelos K. A Blueprint for the Synthesis and Characterisation of Thin Graphene Oxide with Controlled Lateral Dimensions for Biomedicine. 2018. 2D Materials. In submission]. Structural characterisation of these materials, by TEM and AFM (Figure 6.1A), revealed that each material (l-GO, s-GO and us- GO) was composed of sheets of different lateral dimensions as described in Figure 6.1B (further characterisations is provided in Figures 6.S2 and 6.S3). The batch reference for this starting material is GOf3. In summary, l-GO is comprised of sheets with lateral dimension between 1 and 30 μm, whilst s-GO sheets range between 30 nm and 1 μm. The smallest material (us-GO) was characterised by a narrower size distribution, with lateral dimensions ranging from 30 nm to 500 nm. Apart from the lateral dimensions all other parameters, such as oxidation degree and thickness, remained comparable, thereby allowing a lateral dimension dependent study to be fairly undertaken.

In order to track and quantify the presence of GO after instillation, we functionalised the different GO sheets with a chelation agent that could bind to a traceable metallic probe, 111In or 115In. For this we used tetraazacyclododecane-1,4,7,10-tetraacetic acid that was attached to a poly (ethylene)4 glycol molecule which displayed a free amine group (NH2-PEG4-DOTA).

Upon functionalisation with NH2-PEG4-DOTA (mechanism shown in Figure 6.S1) to form

GO-DOTA (we do not include the PEG4 linker in the functionalised material’s name for simplicity), there was a reduction in the sheet lateral dimension, particularly with regard to l- GO-DOTA compared to l-GO, where a 5-fold reduction was observed (Figure 6.1A and B, and Figure 6.S2). The lateral dimensions of s-GO-DOTA and us-GO-DOTA sheets however were still markedly different from l-GO-DOTA and each other. As observed before, the AFM showed that the thickness of all GO materials had slightly increased following DOTA functionalisation45. The size and thickness histogram distributions of the GO-DOTA conjugates can be found in Figure 6.S2. The functionalisation of the GO materials was further studied using Raman spectroscopy and FT-IR. Raman spectroscopy (Figures 6.1B and Figure 6.S3) which revealed that there was no significant change in the I(D) / I(G) ratio following functionalisation. But the FT-IR (Figure 6.S3) evidenced the presence of new bands at 2950 – 2850 cm-1 and an increased band at around 1600 cm-1 as well as the

218

-1 presence of an stronger band between 1260-1330 cm for NH2-PEG4-DOTA compared to the GO samples, which evidences the chemical modification of GO without further defecting the sp2 carbon backbone as suggested by Raman spectroscopy. Throughout our work, we included non-functionalised as controls (batch reference GOf6) which we used to validate the results obtained with GO-DOTA as well as to perform biodegradation studies. The characterisation of this material can be found in Table 6.S1 and is further described elsewhere in detail [Rodrigues A. F., Newman L., Lozano N, Bussy C & Kostarelos K. A Blueprint for the Synthesis and Characterisation of Thin Graphene Oxide with Controlled Lateral Dimensions for Biomedicine. 2018. 2D Materials. In submission].

219

Large GO Small GO Ultra-small-GO GO

500 nm

TEM

DOTA

- GO

5 nm 5 nm 5 nm GO

4 μm 1 μm 1 μm 0 nm 0 nm 0 nm

5 nm 5 nm 5 nm

AFM

DOTA

- GO 4 μm 1 μm 1 μm 0 nm 0 nm 0 nm B

Parameter Technique l-GO s-GO us-GO

Lateral TEM 10 μm [10 – 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] I (D)/I (G) Raman 1.30 0.04 1.35 0.02 1.34 0.03 Parameter Technique 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

Figure 6.1 Physicochemical characterisation of l-GO, s-GO and us-GO before and after DOTA functionalisation.

(A) GO nanomaterials were interrogated, before and after functionalisation, for their morphological and structural features by TEM and AFM. Histogram analysis of their lateral dimensions (TEM and AFM) and thickness (AFM) distributions have been given in Figure 6.S2. The materials spectroscopic features were also analysed with Raman spectroscopy and FT-IR and are shown in Figure 6.S3. (B) A characterisation summary table has also been provided; measured sizes are expressed as means and ranges.

220

Chelation and stability of GO–DOTA[In] conjugates.

The GO-DOTA materials were examined in terms of their ability to chelate metal isotopes - radioactive 111In and non-radioactive 115In. We observed via radio-thin layer chromatography (TLC) that after the chelation reaction using 111In, followed by two washing steps, a small amount of unbound DOTA[111In] (~ 10-13%) was still detectable at the solvent front (Figure 6.2A). However, in comparison to the control DOTA[111In] alone, it was clear that the chelation had occurred. To further characterise the constructs, the absolute quantity of 115In that each GO-DOTA conjugate could chelate was determined using ICP-MS (Figure 6.2B). ICP-MS analysis indicated that all the structures could chelate between 0.2 – 0.4 μmol of In / mg GO-DOTA. This was similar to that reported in our previously publication.45

The stability of the GO-DOTA[115In] constructs was assessed over time in 50% FBS at 37 ⁰C at both 1 and 7 days post chelation (Figure 6.2C). In order to do this, we employed ICP-MS. Similar to what was found using radio-TLC, we observed that there was a small free fraction present in the supernatant that represented unbound 115In. Taking this into account, it was shown that the GO-DOTA[115In] labelled complexes were 75 – 80% stable. At 7 days post chelation, the analysis was repeated and it was found that GO-DOTA[115In] remained stable, with limited further release.

221

A DOTA[111In] functionalised DOTA l-GO s-GO us-GO [111In]

11% 15% 13% 97.5%

Front

Solvent Solvent

oint

P Application Application n.d.

1000 1500 2000 1000 1500 2000 1000 1500 2000 1000 1500 2000 -1 -1 -1 -1

Intensity (a.u.) Intensity Raman Shift (cm ) Raman Shift (cm )Raman Shift (cm ) Raman Shift (cm )

B i) ii)

0.6 100 ) * 75

0.4

mol/mg

 (

C (%) (%) C 50



7 Day 7 Day 1

1 Day 1

7 Day 7 7 Day 7 In ions bound per per bound ions In 0.2

115 25 Stability in 50% FBS, FBS, 50% in Stability 0.0 0

l-GO-DOTA s-GO-DOTA us-GO-DOTA l-GO-DOTA s-GO-DOTA us-GO-DOTA unit mass of GO GO of mass unit 115 115 115 [115In] [115In] [115In]

Amount of of Amount [ In] [ In] [ In]

Figure 6.2 Indium chelation efficiency and stability of GO-DOTA[In] conjugates.

(A) The chelation of Indium (111In or 115In) by GO-DOTA was assessed using radio TLC using 111In with correlative Raman spectroscopy (inset); n.d. = not detectable. (B) i) ICP-MS, where the absolute 115In chelation per mg of GO-DOTA was determined. ii) The stability of the constructs was assessed 24 h in 50% FBS at 37 ºC using ICP-MS. All ICP-MS experiments were completed with at least 3 replicates and statistical differences were assessed using the Kruskal-Wallis test followed by Dunn’s post hoc test. *: p < 0.05.

222

Size-dependent, brain-specific biodistribution of GO following a single intranasal instillation.

The results of the brain-specific biodistribution following a single intranasal administration of 30 μg of l-GO-DOTA[111In], s-GO-DOTA[111In] or us-GO-DOTA[111In] as determined by SPECT/CT imaging, autoradiography and γ-scintigraphy is presented in Figure 6.3. Upon intranasal instillation, a large accumulation of all three radiolabelled GO materials in the nasal cavities was observed (Figure 6.S4). A high signal intensity continued to be detected within the nasal cavities for the entire duration of the experiment, up to 24 h (Figure 6.3A and Figure 6.S4). Raman spectroscopy of the extracted nasal cavities was additionally performed in a separate experiment using non-functionalised materials and confirmed the presence of the materials in the nasal cavities (Figure 6.S5). SPECT / CT evidenced that nose-to-brain translocation occurred after 3 h, although to a limited extent (Figure 6.S4). However 24 h after the intranasal administration, a significant radiation signal was detectable within the region of the brain corresponding to the olfactory bulb for s-GO-DOTA[111In], us- GO-DOTA[111In] and DOTA[111In], but to a far lesser extent for l-GO-DOTA[111In] (Figure 6.3A).

Nose-to-brain translocation of l-GO-DOTA[In 111] although detectable, was not as obvious using SPECT/CT imaging. This was most likely due to the small amounts of material that had undergone translocation. Therefore to collect more revealing evidence, the brains of the animals used for SPECT/CT imaging (2 mice per material) were carefully extracted, fixed and then sectioned into two equivalent sagittal sections; the left half of each brain was then placed onto an autoradiography plate for 7 days. The results of this experiment are shown in Figure 6.3B and demonstrate that all materials including l-GO-DOTA[111In], had translocated to the brain. In correlation with the SPECT / CT, the extent of translocation was least for l- GO-DOTA[111In] and greatest for us-GO-DOTA[111In]. It was noticed that us-GO-DOTA[111In] had translocated to a comparable extent to DOTA[111In], as illustrated by the detection of both probes in many parts of the brain. Such widespread diffusion was not as apparent for mice treated with l-GO-DOTA[111In] or s-GO-DOTA[111In] (Figure 6.3B). It was not clear if out of the us-GO-DOTA[111In] or DOTA[111In] treated samples what material had translocated to the greater extent. This was due to the saturated signal as observed in the autoradiography data that hindered a semi-quantitative analysis.

The SPECT/CT results were further validated by measuring the percentage of the instilled dose of l-GO-DOTA[111In], s-GO-DOTA[111In] or us-GO-DOTA[111In] per gram of micro- dissected brain tissues in a separate experiment using γ-scintigraphy at 24 h post intranasal instillation (Figure 6.3C). The data confirmed the SPECT/CT imaging and autoradiography,

223 and indicated a lateral dimension dependent nose-to-brain translocation of GO-DOTA[111In] was occurring. Overall following, the intranasal administration of 111In labelled GO-DOTA sheets with similar associated radioactive doses, l-GO-DOTA[111In] had translocated to the least extent, while us-GO-DOTA[111In-] had translocated to the greatest extent. The control group on the otherhand, treated with DOTA[111In], also showed a significant accumulation of the probe in the brain. The accumulation of the probe was to a much greater extent than both l-GO-DOTA[111In], s-GO-DOTA[111In] and us-GO-DOTA[111In] treated groups, although there were a high intragroup variation.

224

l-GO-DOTA[111In] s-GO-DOTA[111In] us-GO-DOTA[111In] DOTA[111In]

24 h post h 24 administration

l-GO-DOTA[111In] s-GO-DOTA[111In] us-GO-DOTA[111In] DOTA[111In]

1 cm 3

C l-GO-DOTA[I111In] s-GO-DOTA[111In] us-GO-DOTA[111In] 2 DOTA[111In]

per tissue weight (% / g) / (% weight tissue per Percentage of administered dose dose administered Percentage of 0 4

Cortex StriatumMid Brain 3 Cerebellum Remaining Olfactory Bulb Hippocampus Pons and Medulla 2 Figure 6.3 Brain-specific biodistribution of GO sheets of different lateral dimensions 24 hours following a single intranasal administration. 1 Biodistribution of GO-DOTA[111In] was analysed using (A) SPECT/CT imaging (performed by Dr

Dhifaf Jasim), (B) autoradiography of 0extracted brains and (C) γ-scintigraphy of micro-dissected brain substructures. All scintillation values were subtracted from the background signal. Statistical Cortex StriatumMid Brain significance was assessed using a Kruskal-Wallis test withCerebellum post hocRemaining Dunn’s multiple comparisons Olfactory Bulb Hippocampus test. *: p < 0.05, n = 3 – 4 mice were used per group. Pons and Medulla

225

Raman spectroscopy of brain tissues after 24 h following the instillation of non-functionalised GO was used to confirm the biodistribution results described previously using labelled materials. Aqueously dispersed non-functionalised GO sheets (30 μg) (Table 6.S1) were intranasally instilled to mice, the presence of GO was then screened for in the perfused and harvested brain tissue samples using Raman spectroscopy (Figure 6.4). This approach allowed us to reject the possibility that the radioactive signal used to locate GO sheets in the previous techniques was coming from free DOTA[111In] and not the GO-DOTA[111In] conjugates. Moreover, it helped to validate GO-DOTA as a good system to model the behaviour of non-functionalised GO samples following intranasal administration. The results indicated that us-GO had translocated to all brain regions. In contrast, s-GO was only detected in the olfactory bulbs, pons and medulla, and cerebellum, while l-GO was found in the olfactory bulbs, pons and medulla, but not in other brain regions.

226

Region of the l-GO s-GO us-GO brain analysed

Olfactory Bulb 1000 1500 2000 1000 1500 2000 1000 1500 2000

n.d. n.d. Cortex

1000 1500 2000 .)

n.d. n.d. a.u Hippocampus

1000 1500 2000

n.d. n.d. Striatum

1000 1500 2000 Raman Intensity Intensity Raman ( n.d. n.d. Mid brain

1000 1500 2000 1000 1500 2000

n.d. Cerebellum

1000 1500 2000 1000 1500 2000

Pons and Medulla 1000 1500 2000 1000 1500 2000 1000 1500 2000

Raman Shift (cm-1)

Figure 6.4. Raman-based brain biodistribution of GO sheets of different lateral dimensions 24 hours following a single intranasal administration.

Each Raman spectrum is an average of 9 Raman spectra, each one from a GO Raman positive region found across 3 mice. n.d. = not detectable.

227

Cell-specific brain distribution of us-GO 24 h following a single intranasal instillation.

The qualitative and quantitative results independently suggest that, out of the three sizes of materials investigated, us-GO underwent the most extensive nose-to-brain translocation. In addition, us-GO was identified in all brain regions. It was therefore decided to conduct a separate experiment that focused on the brain-specific localisation and fate of us-GO in two very distant regions (i.e. olfactory bulb and cerebellum) 24 h following a single intranasal administration. The pons and medulla was not chosen due to difficulties in separating it from the spinal cord and various cervical lymph nodes which may interfere with measurements.

To investigate the cellular localisation of us-GO, we cryo-sectioned the brains of treated mice and then performed immunohistochemistry for three major brain cell types, namely neurones (using NeuN immuno-reactivity), astrocytes (using GFAP) and microglia (using IBA-1). To reveal the presence of GO in those specific cells, Raman mapping was conducted on immuno-stained sections. The Raman maps and immunofluorescence images were then superimposed (Raman maps were made 40% transparent), allowing the visualisation of the tissue sub-distribution of us-GO within the brain in a correlative manner. Representative immunohistochemistry images with Raman overlay of sections of the olfactory bulb (Figure 6.5A) and cerebellum (Figure 6.5B) are shown. Due to the limited number of GO Raman positive regions detected within the brain sections, a semi- quantitative representation of the cell-based locations of all detected GO-positive regions (n = 62) found both within the olfactory bulbs and cerebellums of 3 different animals is presented as a pie chart (Figure 6.5C). The data indicates that us-GO sheets were predominantly found in close association with microglia that were positively identified by IBA- 1 immuno-staining in both brain regions (Figure 6.5C).

228

Immunofluorescence - Raman correlation Raman map A 1x 12x 40x only 80 % 500 μm

(Microglia) (Microglia)

– 40 % 50 μm 25 μm 25 μm IHC 500 μm 80 %

(Astrocytes) (Astrocytes) 40 % – 50 μm 25 μm 25 μm

IHC 500 μm 80 %

(Neurons) (Neurons) – 40 %

50 μm 25 μm 25 μm IHC IHC

Immunofluorescence - Raman correlation Raman map B 1x 12x 40x only

500 μm 80 %

(Microglia) (Microglia) 5 μm – 40 %

50 μm 25 μm 25 μm IHC IHC

500 μm 80 %

(Astrocytes) (Astrocytes) – 40 %

50 μm 25 μm 25 μm IHC IHC

500 μm 80 %

(Neurons) – 50 μm 25 μm 2525 μ mμm 40 % IHC IHC

Figure 6.5 Qualitative and semi-quantitative location of GO 24 h following a single intranasal administration.

Representative fluorescence images of brain cells stained green (IBA-1 positive microglia, GFAP positive astrocytes or NeuN positive neurones, counterstained with DAPI, indicated by the blue colour) with Raman overlay (40% transparency) of the (A) olfactory bulb and (B) cerebellum of treated mice are shown. Magnifications of regions of interest are provided (12x and 40x) that highlight the location of GO as detected by overlaid 50% transparent Raman map. The respective non-transparent 40x magnified Raman maps have been provided for comparison with a heat scale bar referring to the percentage of similarity of the Raman positive region to a GO reference spectrum which has been given in Figure 6.S6. Bright field images showing the regions of interest that underwent Raman mapping taken directly from the Raman mapping software are shown in Figure 6.S7 for reference. (C) All detected GO positive regions (n = 62) were scrutinised in terms of their relative position to cells, allowing a semi-quantitative distribution of GO tissue location as expressed in the pie chart.

229

Time-dependent distribution of us-GO at 1 and 7 days following a single intranasal instillation.

To quantitatively assess the brain biodistribution of us-GO at different time points and after 24 h, we could not use γ-scintigraphy due to the short radioactive half-life of 111In, which could hamper reliable detection at the latter time point. We therefore used the cold 115In isotope, chelated it with the us-GO-DOTA construct, and then measured the brain biodistribution of us-GO-DOTA[115In] by ICP-MS at Days 1 and 7 after a single intranasal administration. The results are presented in Figures 6.6 (% administered dose per dry weight of tissue) and Figure 6.S8 (% administered dose per whole tissue). The results of this experiment confirmed that of the γ-scintigraphy (Figure 6.3C), which strengthened the conclusions concerning the widespread nose-to-brain translocation of us-GO. After 7 days, the amount of detected us-GO-DOTA[115In] probe remained constant overall, however there were some non-significant variations when comparing the individual micro-dissected brain regions at days 1 and 7. The DOTA[115In] control exhibited a similar qualitative biodistribution to us-GO-DOTA[115In], although the initial translocation to the olfactory bulb was greater than for us-GO-DOTA[115In] (Figure 6.3C). After 7 days and in contrast to us- GO-DOTA[115In], the vast majority of the DOTA[115In] cleared from the brain, with only some traces still detectable within the cortex, hippocampus and cerebellum.

230

Day 1 (us-GO-DOTA[115In]) Day 7 (us-GO-DOTA[115In]) Day 1 (DOTA[115In]) Day 7 (DOTA[115In])

3 pertissue weight (% g) / 0 6

Percentageof Administered dose Cortex Striatum Midbrain Cerebellum Remaining Olfactory Bulb Hippocampus Pons and Medulla

Figure 6.6 Evolution of the biodistribution profile of us-GO at days 1 and 7 following a single 3 intranasal administration, assessed by ICP-MS.

The biodistribution profile of us-GO-DOTA[115In] at Days 1 and 7 was determined following a single intranasal administration. Each micro-dissected brain region was analysed with ICP-MS for the presence of 115In. The data has been expressed as percentages of the administered dose per weight (g) of dry tissue and subtracted from the brain regions of Dex 5% treated animals. Data corresponding 0 to the percentage of administered dose per anatomical brain region is presented in Figure 6.S8.

Statistical significanceCortex was assessed using a Kruskal-Wallis multiple comparisons test with Dunn’s Striatum Midbrain post hoc test. *: p < 0.05, n = 4 mouseCerebellum were used perRemaining group. Olfactory Bulb Hippocampus Pons and Medulla To confirm the ICP-MS data by a technique not relying on a probe (i.e. 115In) but directly on the GO sheet’s intrinsic properties, we administered non-functionalised us-GO as before and used Raman spectroscopy on the perfused and extracted brain sample homogenates (see experimental section for details). We obtained a qualitative Raman-based biodistribution at days 1 and 7 following intranasal administration of us-GO sheets (Figure 6.7). After 1 day, us-GO could be found in all brain regions as previously demonstrated (Figures 6.3, 6.4 and 6.6). After 7 days, however, us-GO could only be detected within the olfactory bulb and cerebellum, in good agreement with the ICP-MS data (Figure 6.6). Some differences however were apparent such as the presence of us-GO in the pons and medulla at day 7, which is not apparent from the ICP-MS based biodistribution.

Overall, these results (Figures 6.6 and 6.7) revealed that the distribution of us-GO in the brain following intranasal administration is a dynamic process.

231

us-GO us-GO Region of the (Day 1) (Day 7) brain analysed

Olfactory Bulb 1000 1500 2000 1000 1500 2000

n.d. Cortex 1000 1500 2000 1000 1500 2000

.) n.d.

a.u Hippocampus

1000 1500 2000

n.d. Striatum

1000 1500 2000

Raman Intensity Raman Intensity ( n.d. Mid brain

1000 1500 2000

Cerebellum

1000 1500 2000 1000 1500 2000 Pons and Medulla

1000 1500 2000 1000 1500 2000 Raman Shift (cm-1)

Figure 6.7 Evolution of the biodistribution profile of us-GO at day 1 and 7 following a single intranasal administration, assessed by Raman spectroscopy.

The presence of GO was confirmed in different micro-dissected brain regions using Raman spectroscopy. Each Raman spectrum is a collative of 9 GO Raman positive regions found across 3 mice. n.d. = not detectable.

232

The in vivo biodegradation of nose-to-brain translocated us-GO after a single intranasal instillation.

To explore whether in vivo biodegradative processes were at least partially responsible for the reduction in the amount of us-GO after 7 days, we conducted a more refined analysis of the GO Raman spectra to interrogate the crystalline state of the us-GO sheets over time (Figure 6.8). We focused this experiment on the analysis of the materials present within the olfactory bulbs, where our data indicate the majority of the material was residing, and the cerebellum which serves to represent an anatomical structure where the materials had further translocated as evidence both using Raman spectroscopy and the various imaging methods. We extended the time line of the experiment to 1 month in order to assess changes in the Raman spectra more confidently. We were able to reproducibly detect GO at the 3 different time points (i.e. 1, 7 and 28 days) in both the olfactory bulb and cerebellum of the treated mice (Figure 6.8A). There was a decrease in the GO Raman signal intensity over time and at the 1 month time point the signal was at its minimum with respect to both brain regions. The rate at which the intensity decreased was greater for the materials in the olfactory bulbs than those in the cerebellum. After determining the I(D) / I(G) ratio at each time point, we saw an initial rise of the ratio in the first seven days in the olfactory bulb, whereas it remained largely unchanged within the cerebellum (Figure 6.8B). At the 1 month time point, however, there was a statistically significant decrease in the ratio. This was in parallel to a reduction in the signal intensity compared to day 7 for both brain regions. These results strongly support the idea that reduction in the amount of us-GO in brain regions could be ascribed at least partially to biodegradation.

233

A Olfactory Bulb Cerebellum )

15 ) 15

a.u.

3 x10

( 10 10 x10 Day 1 ( Day 7 5 Day 28 5

0 0 1000 1500 2000 1000 1500 2000

Raman Intensity Raman -1 -1

Raman Shift (cm ) Intensity Raman Raman Shift (cm )

B Olfactory Bulb Cerebellum

2.5 * 2.5 ** 2.0 2.0 15 **

1.5 1.5 I(D)/ I(G) I(D)/

10 1.0 1.0 I(D) / I(G) / I(D)

0.5 0.5 1 7 28 5 1 7 28 Time (Days) Time (Days)

0Figure 6.8 Biodegradation of us-GO over 28 days following a single intranasal administration. 1000 1500 2000 The olfactory bulbs (left) and cerebellum (right) of mice treated with us-GO were isolated, physically homogenised and screened for the presence of GO using Raman spectra at days 1, 7 and 28. (A) The average Raman signal and (B) the box plots of the I(D) / I(G) at each time point are given. At each time point, n = 5 spectra were collected from each of the 3 mice initially treated. Statistical significance was assessed using a Kruskal-Wallis test with Dunn’s post hoc test. *: p < 0.05, **: p < 0.01.

Translocation and biodegradation of us-GO in the brain after multiple intranasal instillations.

We further assessed whether the biology of the brain would be able to handle us-GO in the same way, if the same dose was administered multiple times (5 doses over 5 consecutive days, 1 dose/day), potentially allowing an increase in the amount of translocated us-GO

234

(Figure 6.S9). We found that there was only a modest increase in the number of GO positive cells per mm2 of brain tissue in both the olfactory bulbs and cerebellum following the repeated exposures, in comparison to a single administration (Figure 6.S9). Despite a 5- time repeated exposure to 30 μg of us-GO, there was not a 5-time increase in the amount of GO found in the brain. When analysing the collected Raman spectra, we observed a decrease in the Raman intensity over time and a characteristic decrease in the I(D) / I(G) ratio as well as the signal intensity from day 7 to 28 (Figure 6.S10), confirming the biodegradation observed after a single administration. Overall, these results suggest that translocation from nose-to-brain of us-GO was not a cumulative process. They also support the idea that biodegradation is a constitutive process, not impaired by repeated exposures, that could play a key role in the elimination of translocated materials regardless of the chronicity of exposure. In all of our studies in correlation with the analysis of the GO materials, we continuously measured the weights of all the mice during our experiments. The results showed that all mice displayed characteristic growth patterns relative to aged matched Dex 5 % treated controls (Figure 6.S11).

6.5. Discussion.

The traditional notion that the lungs and other major components of the respiratory tract such as the trachea are targets of inhaled particles has held strong. In recent times however, evidences from various epidemiologic studies46, clinical trials47 as well as animal experiments20 have demonstrated that extra-pulmonary organs can also be affected. For instance, the CNS has been shown to be a target of inhaled nanoparticles including graphite-derived ultrafine carbon particles22 and various metallic nanoparticles.48 In the present study, the potential of GO sheets of different controlled lateral dimensions to undergo nose-to-brain translocation following intranasal instillation was explored. The fate of translocated GO materials in the brain using imaging, and spectrometric and spectroscopic techniques was also analysed.

Characterisation of materials.

In order to perform these different experiments, both non-functionalised and DOTA functionalised materials were used. The purpose of DOTA functionalised materials was to allow labelling of the GO sheets with metallic probes that could facilitate their identification and quantification within tissues. We first functionalised thin GO sheets (single to few layers) of controlled lateral dimensions44 [Rodrigues A. F., Newman L., Lozano N, Bussy C &

235

Kostarelos K. A Blueprint for the Synthesis and Characterisation of Thin Graphene Oxide with Controlled Lateral Dimensions for Biomedicine. 2018. 2D Materials. In submission] with

DOTA connected to a PEG4 linker group (NH2-PEG4-DOTA) via an epoxide ring-opening reaction.49 This strategy allowed the functionalisation of the GO materials without further perturbing the graphitic backbone or causing a chemical reduction of the GO sheets. We have previously shown that this epoxide ring-opening reaction is a versatile means to functionalise GBMs49 [Orecchioni et al. 2017]. Following functionalisation, the sheet-like morphology of all the materials following functionalisation was maintained. However, the lateral dimensions of the l-GO-DOTA sheets were smaller than that of the starting l-GO sheets. This difference was not as obvious for s-GO or us-GO and their respective functionalised derivatives. Nevertheless, the lateral dimensions of l-GO-DOTA remained greatly in excess of those of the s-GO-DOTA and us-GO-DOTA constructs, still permitting a lateral dimension based examination of the nose-to-brain translocation of GO. The observation of a reduction of the lateral dimensions of GO sheets following chemical functionalisation has been previously reported by us50 as well as other groups.51

The reduction in l-GO sheet lateral dimension could be a consequence of the cooperative alignment of epoxy groups, which is expected to effectively create fracture lines in the 52, 53 sheets on the l-GO sheets. Upon epoxide ring-opening using the amine of the NH2-

PEG4-DOTA group as the nucleophile, a mass unzipping of the sheet along these epoxide- related fracture lines could occur, which would result in smaller graphene sheets. Since this phenomenon was not observed to the same extent for s-GO or us-GO sheets, we hypothesise that the cooperative alignment of epoxy groups is not as prominent. Most likely because the sonication of l-GO, required for the acquisitions of these smaller sheet sizes, had already resulted in an unzipping processes. This means that the previously aligned epoxides on the l-GO sheets now exist as more stable configurations on the edges of s-GO and us-GO, such as semiquinones.53 Consequently, epoxide ring-opening reactions may involve the reactive epoxides that still remain on the basal planes of the more numerous smaller sheets.

It was observed that the sheet thickness had increased on average for all materials. We50 and others51, 54 have previously reported this observation following the functionalisation of GO with moieties including DOTA derivative bearing an isothiocyanate moiety50, dextran55, or polyethyelene glycol.51 This observation may have been due to the added DOTA functionalities which could further increase the effective thickness of the sheets. The graphitic backbone of the materials was interrogated using Raman spectroscopy, which evidenced the presence of the characteristic D and G scatter bands (1330 cm-1 and 1595

236 cm-1). The spectra of the materials were then scrutinised for their I(D) / I(G) ratio, which is a commonly used parameter to assess disorder within GBMs.56 When comparing I(D) / I(G) of GO sheets and GO-DOTA conjugates, the value of the ratio did not increase significantly. This was expected since the conditions used for the epoxide ring-opening reaction have previously been shown not to further defect the graphitic lattice.49

FT-IR analysis revealed differences between the non-functionalised and DOTA functionalised GO sheets, evidencing the chemical modification. Importantly, for the GO- DOTA sheets, there was an enhanced and broader band near 1600 cm-1 and between 1260- 1330 cm-1 in the region where the amide C=O stretch, and the C-N stretch and C-H bending are expected respectively, compared to the non-functionalised GO. Moreover for GO-DOTA compared to GO, there were 2 new bands at 2850 - 2970 cm-1, which can be attributed to the stretching of the aliphatic C-H groups, this supports the successful functionalisation of

GO with NH2-PEG4-DOTA, which will result in a product that contains N-H, C-N bonds and is rich in aliphatic C-H groups. Similar observations have been previously reported.45, 57

We then investigated the ability of our materials to successfully chelate indium isotopes (111In or 115In) and label GO. In doing so, the successful chemical modification of GO with

NH2PEG4-DOTA was confirmed and ensured that the labelling of our materials was suitable for biodistribution studies (Figure 6.2). With radio TLC, we demonstrated the successful retention of 111In by GO-DOTA. Using ICP-MS, we show that all 3 GO-DOTA materials chelated a comparable and significant amount of 115In, allowing efficient labelling. The GO- DOTA[115In] structures were shown to be approximately 75 – 80 % stable when incubated in serum at 37 ⁰C over 24 h. These values compare to those reported in previous studies.58, 59 No major significant decreases in stability were detected for up to 7 days, indicating the stability of our materials was maintained. The initial release is likely due to a small amount of unreacted NH2PEG4-DOTA, which may have been physisorbed onto the surface of the GO and was gradually released. The high binding and stability of these [In]-labelled GO-DOTA structures is different from the labelling of non-functionalised GO. We have previously investigated and shown that non-functionalised GO sheets labelled via simple mixing with indium results in structures that exhibited poor stability in serum over time at 37 ⁰C, where the majority of bound indium is released in 24 h.50 Taken together, the physicochemical and labelling characterisations evidence the successful synthesis of GO-DOTA sheets of controlled lateral dimensions to be used in our biodistribution experiments.

Apart from the presented labelling method, other strategies have been also reported such as with radioactive iodine60, which requires no chelating agent based attachment. These strategies however employ highly oxidising species that can significantly damage the

237 materials graphitic backbone. Moreover, the physiological accumulation of iodine within the thyroid gland can lead to difficulties in interpreting data.55 Physical adsorption strategies to attach radioactive metals, bound within a chelating agent, onto the surface of GO via π-π interactions have also been investigated. These strategies however can result in unwanted large increases in sheet thickness61, changing the behaviour of the materials and therefore their validity as model systems to approximate the biodistribution of single to few-layer GO sheets.62, 63

Brain-specific size-dependent biodistribution of GO following single intranasal instillation.

After the successful preparation of labelled GO constructs with 3 different average lateral dimensions, we started to investigate their respective brain biodistribution after intranasal administration. The SPECT/CT imaging data revealed that during the early time points (i.e. 30 min after instillation) there was no convincing evidence of transfer of materials across the cribriform plate. The cribriform plate is a sieve-like structure that separates the CNS from the roof of the nasal cavities. After 3 hours, a weak signal could be seen within the brain remit, providing evidence for translocation of us-GO-DOTA[111In] and DOTA[111In] to the CNS, specifically via the olfactory bulb (Figure 6.S4). Other studies investigating nose-to-brain transport of nanoparticles, such as quantum dots, have also observed a short delay from the time of administration until there is some evidence of translocation to the brain64, this delay may be due to the physical size of the material which limits their diffusion. Twenty four hours following the administration, evidence of translocation to the olfactory bulb was however clearly apparent for two of the conjugates, namely s-GO and us-GO, and the DOTA control (Figure 6.3A).

The initial detection of materials within the olfactory bulb 3 h after a single administration suggests it is unlikely that the transport occurs only via axonal transport. In this pathway, materials in the nasal cavities are up taken by neurones via endocytosis or pinocytosis and are then actively transported through the cribriform plate in a retrograde fashion towards the olfactory bulb in the brain.19, 30 Axonal transport has been shown to be relatively slow. Several studies report that up to 24 h are required before any material reached the CNS via axonal transport.65 Therefore it could be in part responsible for some translocation of materials into the brain, particularly at the 24 h time point, but it does not explain the initial rapid transport observed for us-GO within 3 h. Another possibility is the transport of materials from their initial locations after intranasal administration (i.e. nasal cavity, respiratory tract,

238 stomach, intestine) into the systemic circulation, from which they would then cross the BBB and migrate into the CNS.31 This mechanism however is less likely due to the initial accumulation of materials within the olfactory bulb observed here, which would not be expected via the systemic/BBB route. Other mechanisms of transport include transcellular conveyance of the materials through the olfactory sustentacular epithelial cells or even paracellular transport through the narrow tight junctions between these sustentacular epithelial cells.15 These transport mechanisms can occur continuously, over much shorter time frames, and are more likely to explain the rapid appearance of material within the olfactory bulb. Other intranasally administered materials including drug molecules, fluorescence molecular tracers, exosomes as well as quantum dots have also been reported to be translocated to the brain within short time frames.25, 31, 64, 66-68

In a complementary fashion, the SPECT/CT imaging (Figure 6.3A and Figure 6.S4), the autoradiography maps (Figure 6.3B) and the γ-scintigraphy measurements (Figure 6.3C), together with the Raman-based biodistribution (Figure 6.4) provide quantitative and qualitative evidences to support the size-dependent nose-to-brain translocation of GO sheets. This was interesting and could be explained by size restriction issues. For instance, De Lorenzo determined in 2-month old rabbits that the axons of olfactory neurones, which could in part be responsible for nose-to-brain translocation, had an approximate diameter of 200 nm; he also noted that other axons smaller than 100 nm were present.19 Such a size restriction would therefore only allow transport of materials that did not exceed the diameter of the neuronal axons. Transcellular transport through the olfactory sustentacular epithelial cells is also possible but again will be size restricted. Paracellular diffusion pathways, between the olfactory sustentacular cells and neurones, are even more size limited since the tight junctions between cells of the olfactory mucosa are reported to be 3.9 - 8.4 Å, allowing negligible transport of materials greater than 15 Å.69 Even in the presence of efficient absorption enhancers, these tight junctions will only open by 15 - 20 times more70 which is far below the nominal lateral dimensions identified during characterisation of the tested GO materials. However, the frequent turnover of the olfactory neurones may provide some infrequent opportunities for larger particles to penetrate the barriers.71, 72 Overall, it makes sense that only the smallest and thinnest of structures would be able to undergo nose-to- brain transport via these routes. Other studies have also reported that size plays a critical role in determining the extent to which nanoparticles undergo nose-to-brain translocation25. This finding could have a major relevance for companies that aim to produce water-stable 2D materials with safer (by design) features.73 Towards this aim, materials with lateral dimensions above 1 μm would appear safer in terms of translocation to the brain in an inhalation exposure scenario.

239

As per the ICRP model14, most of the materials are however expected to remain lodged within the nasal cavities after an intranasal administration.74, 75 To confirm this expectation, we harvested the nasal cavities from animals treated with non-functionalised GO 24 h after instillation. Black materials were found in the tissue (Figure 6.S5). Surprisingly, the amount of remaining materials found in the nasal cavities after 24 h was size-dependent, with the lowest amount for l-GO treated animals and the highest for us-GO treated animals. This suggests that the size-dependent translocation of GO to the brain, which follows the same trend, could be also partly explained by the retention time of each GO material in the nasal cavities. A Raman-based analysis was further performed to confirm the graphitic nature of these black materials. At 24 h post administration, the persistence of all three types of GO in the nasal cavities was identified (Figure 6.S5B).

The evidence also suggests that after 24 hours, besides being present within the olfactory bulbs (Figure 6.3), small quantities of us-GO were able to further translocate to more distant structures such as the cortex, striatum, hippocampus, midbrain, cerebellum, as well as the pons and medulla (Figures 6.3C, 6.4, 6.6 and 6.7). Previous studies have also reported the presence of materials in more distant brain structures following the intranasal administration of carbon particles22, iron oxide nanoparticles48 and exosomes.25 On close analysis of the sagittal autoradiography maps (Figure 6.3B), it can be seen that us-GO materials appear to reach these distal brain structures after passing through the olfactory bulb (as suggested by the gradient from the olfactory bulb to distal regions) rather than being introduced by an alternative means. Interstitial extra neuronal transport pathways (i.e. in between cells) may explain the diffusion towards distal regions. These pathways are expected to directly transport us-GO sheets from the nasal cavities, into the cerebrospinal fluid especially within the subarachnoid spaces that surrounds the brain.31 The materials could subsequently move toward deeper sites of the brain by convective flow, driven by arterial pulsations76 as they translocate through the perivascular spaces that surround surface and penetrating arteries.68, 77 It should be noted however that the total quantity of GO that reached the brain was extremely low, on average ~0.012% of the initial dose in total (Figure 6.S8). Nevertheless, the dose of the material per gram of dry tissue in the olfactory bulb was higher than other studies using nanoparticles such as 20 nm silver nanoparticles20, and comparable to studies using similar materials such as ultrafine carbon particles derived from graphite, which may well entail allotropes of sp2 hybridised carbon including graphene.22 But it was far lower compared to the translocation values reported in studies that explore the use of nanoparticles for nose-to-brain drug delivery.78 Based on the distribution of us-GO that was found in the brain, it is unlikely that materials had entered via the systemic circulation upon

240 crossing the BBB. Moreover, previous intravenous biodistribution studies conducted by us which used even higher doses of GO show that the majority of the materials underwent extensive urinary excretion, while the remnant fraction was taken up by components of the mononuclear phagocytic system such as the spleen, with no obvious uptake into the brain.50, 79 Furthermore, previous literature studies that looked specifically at the prospect of transfer of similarly sized GO sheets from the systemic circulation to the brain showed that, even with exceedingly high dosages of intravenously administered GO, there was negligible evidence of such transport across the BBB.80, 81 Along with the live imaging data, this evidence supports the concept that the accumulation of GO materials within the brain after an intranasal administration occurred primarily via nose-to-brain translocation using an interstitial extra neuronal transport pathway. Further investigations will however be required to confirm the translocation mechanisms.

Cell-specific location of us-GO.

Using Raman map analysis in correlation with immunofluorescence microscopy, we wanted to investigate more precisely the specific location of us-GO sheets within the olfactory bulb and the cerebellum, selected for their distance apart, with respect to three main cell types of the brain (Figure 6.5). In both brain regions, the majority of the materials were closely associated with microglial cells (IBA-1 positive), with smaller fractions associated with astrocytes (GFAP positive cells) or neurones (NeuN positive cells). This was significant, as it suggested that following intranasal instillation, us-GO sheets were able to translocate to the deep brain parenchyma, albeit in diminutive quantities, and interact with the different cell populations of the brain. Preferential interactions of us-GO with microglial cells in comparison to other cell types was expected, as microglial cells are the resident macrophages of the CNS, and should be able to recognise us-GO sheets as foreign materials.82, 83 Sequestration by microglial cells after intranasal administration and nose-to- brain transport has been previously reported for various nanomaterials such as exosomes25, as well as silver nanoparticles and quantum dots.20, 64 As resident macrophages of the CNS, microglial cells are the main representative cells of the immune system. Under certain conditions, internalisation of nanomaterials by macrophages has been associated with cytotoxic and inflammatory responses.84, 85 Therefore, further research is needed to assess the potential impact of the observed association of GO materials in microglial cells after intranasal administration, in particular on the brain physiology and functions.

241

Changes in the biodistribution profile over time.

Following the initial entry of GO materials within the brain, both ICP-MS (detecting 115In) and γ-scintigraphy (detecting 111In) provided similar brain regional distributions for the 1 day time point, with the highest accumulation of us-GO occurring in the olfactory bulb. When examining the 7 day time point, we noticed that the overall percentage of the administered dose of us-GO materials present in the total brain had remained on average constant unlike the DOTA[115In] control which had almost completely cleared from all brain regions by day 7 when considering the total measured 115In content in all brain regions at day 1 (0.023%) and day 7 (0.0005%). There were some exceptions however such as the cerebellum for instance where the amount had non-statistically increased (Figure 6.S8). This indicated that us-GO sheets and DOTA demonstrated different evolutions of their distribution profiles, with a greater retention of us-GO in several brain regions. The initially similar biodistribution profile supported the idea that us-GO had travelled mostly though the interstitial space, where small molecules such as DOTA are likely to be, allowing the diffusion from the olfactory bulb to distal regions (i.e. cerebellum) over time. It is possible that after 7 days the DOTA[115In] had been cleared by mechanism such as lymphatic transport 77, 86, 87, especially relevant to materials that were in the interstitial space and not entrapped in cells, the fate of the us-GO sheets however was not as clear. A clear question however was could the minor amount of translocated and retained usGO sheets be biodegraded within the brain. This was a particularly intriguing concept, since the majority of the material has been shown to be associated with microglial cells which are known to remove and degrade unwanted materials.

In vivo biodegradation of brain localised us-GO over the course of a month.

In a previous study [Chapter 5], we have shown that in vivo degradation of GO sheets is plausible, especially when materials interact with cells that have the ability to perform degradation such as macrophages.39, 88 We previously demonstrated that intravenously administered GO sheets were predominantly sequestered within a subpopulation of specific macrophages in the spleen, named marginal zone macrophages [Chapter 6]. We also revealed that these splenic macrophages which entrapped GO sheets mediated their structural degradation over the course of 9 months. In the past, we have conducted several studies with functionalised carbon nanotubes and demonstrated both in vivo37 and in vitro40, 42 that microglial cells, the macrophages of the brain, are able to bio-degrade CNMs. These results are in turn supported by a whole host of other degradation studies, which demonstrated the capability of mammalian enzymes to degrade graphene-based materials as recently reviewed.89, 90

242

Based on this knowledge and the identification of the preferential interaction of us-GO with microglial cells (Figure 6.5), we used Raman spectroscopy to probe the crystalline state of the GO sheets56 after their translocation into the brain (Figure 6.8). The timeline of this part of the study was extended to 1 month post instillation, allowing the state of the materials to be examined at a distant time point from the time of administration, taking into consideration the kinetics of degradation observed in other tissues in our previous studies [Chapter 6]. Within the olfactory bulb, the GO Raman signal intensity and integrity was found to gradually decrease (Figure 6.8A). In addition to the intensity, the changes in the associated I(D) / I(G) ratio were also assessed. The I(D) / I(G) increased from 1 to the 7 day time point, indicating that the degree of defects had increased within the materials56 (Figure 6.8B). At the 1 month time point, the I(D) / I(G) ratio decrease to a level below that of the starting materials. This signifies the further increase in defects and amorphicity within the 2D graphene structures56 as predicted by the Tuinstra-Koenin equation and hence its biodegradation. The continuation of such processes is expected to cause the gradual attenuation of all the component peaks associated with the Raman spectra of the GO, and may explain in part the gradual decrease in GO signal intensity and integrity.56, 91 Overall, these results suggest that us-GO sheets were degraded in the brain after they had been in contact with microglial cells (Figure 6.5). This is in agreement with different studies on the biodegradation of oxidised sp2 hybridised carbon nanomaterials conducted by us and others.37, 40-42, 92 The mechanisms by which microglial cells degrade carbon nanomaterials require further investigation, but may involve peroxidase family enzymes or NADPH oxidase or other cellular components, as reported for macrophages and neutrophils.39, 93, 94 The corresponding results from the cerebellum indicate that the GO materials undergo a similar transformation in this brain region, albeit at a slower, more gradual rate.

From a toxicological perspective, evidencing that microglial cells may be able to degrade thee small quantity of us-GO sheets that had translocated to the brain was a relieving observation. It suggests that, even though us-GO sheets can enter the brain from the nasal cavities, their residence may not be permanent, reducing the potential risk of long-term damages to the tissue due to material persistence. To challenge the robustness of this statement, we conducted a further multiple exposure experiment (5 repeats of an equivalent dose administered in the single exposure study, administered 1 dose/day over 5 consecutive days). Surprisingly, only a modest increase in the amount of materials deposited within the brain was identified (Figure 6.S9). This slightly greater presence of GO decreased over the course of a month in a similar fashion to what was found after single exposure. However, at the 1 month time point, the GO Raman signal intensity was greater following the repeated exposure than the single exposure protocol (Figure 6.S10).

243

Overall these results suggest that the brain translocation of 2D sheets following nasal instillations of GO suspensions under the conditions tested here is low and occurs in a size- dependent manner. Also, the translocated GO materials undergo degradation within the brain, which should limit their potential impact. Previous studies have shown that the degradation products of CNMs, derived in vitro via enzymes, did not induce enhanced toxic effects in vivo following their pulmonary and peritoneal administration.93, 95 However, further in depth studies will be required to assess whether or not these trace amounts of materials and their degradation by-products (possibly highly amorphous oxidised materials and oxidised aromatic structures) can impair the brain functions and physiology. Throughout our experiments, we regularly measured the weights and observed the general well-being and body condition of all treated animals, in agreement with standards defined previously.96 No abnormal weight loss (Figure 6.S11) or behaviours were observed relative to age-matched controls treated with dex 5 %. These observations support the lack of detrimental effects of the translocated materials and related biodegradation by-products, although longer term studies with repeated exposures should confirm this statement.

Conclusion.

This study provides evidence to suggest that, following the intranasal administration of aqueously dispersed GO sheets, the materials undergo a size-dependent translocation to the brain. The smallest sheet size category (us-GO, 30 – 500 nm) underwent the greatest translocation to every examined brain regions, but most notably to the olfactory bulb. The actual percentage of materials that translocate to the brain, however, is a diminutive fraction of the administered dose. Focusing on us-GO, we observed that a large fraction of the translocated materials resided in close association with microglial cells. Over the course of a month, the amount of us-GO sheets present within the brain undergo an in vivo biodegradative processes. This occurs most likely for the fraction of materials that were entrapped in microglial cells. To the best of our knowledge, the size-dependent nose-to-brain translocation of GO sheets and their subsequent in vivo biodegradation has not been demonstrated before. This study adds to the growing awareness of the potential impact of the biological interactions of carbon nanomaterials with biological systems, which is critical considering the predicted increased technological utility of these materials in coming years.

244

6.6. Experimental.

Synthesis of non-functionalised GO sheets of differing lateral dimensions.

The synthesis of GO of differing lateral dimensions was in accordance with that described in previous reports, including production under a non-pyrogenic environment 93 [Rodrigues A. F., Newman L., Lozano N, Bussy C & Kostarelos K. A Blueprint for the Synthesis and Characterisation of Thin Graphene Oxide with Controlled Lateral Dimensions for Biomedicine. 2018. 2D Materials. In submission]. During the procedure a multiplicity of measures were taken to minimise the contamination of GO with endotoxin, which included the use of depyrogenated glassware and sterile water for injections and ensuring all post- synthesis handling of GO dispersions was carried out under sterile laminar flow hoods. The mass concentrations of materials were determined by lyophilising a known volume of the synthesised aqueously dispersed GO material. The dried material was then weighed allowing the mass concentration of the produced material to be known. All concentrations were 2.0 mg/mL.

PEG4-DOTA functionalisation of GO sheets.

NH2-PEG4-DOTA (9 mg, 0.0129 mmol) was added to an aqueous suspension of GO starting material (us-GO, s-GO or l-GO) (9 mg; 1 mg/ml). The mixture was left to react for 2 day under continuous stirring at room temperature. GO-DOTA was then directly dialysed in MilliQ water for 4 days to remove free unreacted species. The final dialysed GO-DOTA was stored in water without further treatment.

Metal isotopic labelling of DOTA functionalised GO sheets.

Each GO-DOTA dispersion was diluted with an equal volume of 0.2 M ammonium acetate 111 115 buffer pH 5.5, to which 2 – 20 MBq as InCl3 (radioisotope labelling) or 10 μg cold InCl3 (isotope labelling) was added. For the DOTA chelator alone, the equivalent radioactivity dose (was added for 111Incl3) for radiation based bioimaging experiments, or the average 115 amount of InCl3 bound by the three GO-DOTA conjugates as determined via ICP-MS for mass spectrometry based analysis. The indium was left to react with the GO-DOTA for 60 min at 60 ⁰C under agitation, after which the reaction was quenched by the addition of 50 μL of a 0.1 M EDTA chelating solution. The dispersion was then centrifuged at 13,000 rpm for 111 115 30 minutes to form a GO- PEG4-DOTA[ In or In] pellet.

When using 111In, the supernatant separated from the pellet and both were measured for the radioactive emission, the supernatant was then discarded. The pellet was then re-

245 suspended in dextrose 5% aqueous solution ready for use. All unwanted or waste 111In containing supernatants were aliquoted into Eppendorfs and left to decay for 9 half-lives in specified lead containers. After this the liquid was transferred into a sealed waste container, that was then disposed appropriately.

When using 115In, the supernatant were discarded. The pellets were re-dispersed and underwent centrifugation again as a further washing step. The supernatant was treated as described above and the pellet was made up to 30 μg/mL using dextrose 5% aqueous solution ready for use. All unwanted 115In containing supernatants were placed into a sealed waste container that was then disposed appropriately.

Transmission electron microscopy.

A 10 µL aliquot of sample (200 µg/mL) was placed on a glow-discharged carbon-coated copper grid (CF400-Cu) (Electron Microscopy Services, UK). Filter paper (Merck-Millipore, UK) was used to absorb the excess unbound material. Then samples were observed with a FEI Tecnai 12 BioTWIN microscope (Techni, Netherlands) utilising an acceleration voltage of 100 kV. Images were captured with a Gatan Orius SC1000 CCD camera (GATAN, UK).

Atomic force microscopy.

Sample preparation was completed on freshly cleaved mica, treated with 40 μL of poly-L- Lysine (Sigma-Aldrich, UK) to present a positively charged surface, allowing adhesion of the GO sheets. Aliquots of 10 μL of the respective solutions were then transferred onto the mica surface coated with poly-L-lysine and left to adsorb for 2 min. Unbound GO sheets were then removed via gentle washing with 2 mL of MilliQ H2O and left to dry at room temperature. During analysis, a multimode atomic force microscope (Bruker, UK) was used in tapping mode in order to reduce damage to the samples for height (trace and retrace) and amplitude. Scans were completed using an Otespa tapping mode tip (Bruker, UK) using the following parameters: a scan rate of 1 Hz; lines per scan of 512; an integral gain of 1 and a proportional gain of 5; an amplitude set point value of 150 mV was maintained constant between all measurements. Scan areas were set at 2500 μm2, 400 μm2, 100 μm2, and 25 μm2. Post image processing was completed using the Bruker Nanoscope Analysis Version

1.4 software (Bruker, UK).

Raman Spectroscopy.

Samples were prepared for analysis via drop casting 20 μL of GO (100 µg/mL) dispersion onto to a glass slide. Samples were left to dry for at least 2 h at 37 °C. Spectra were

246 acquired using a micro-Raman spectrometer (Thermo Scientific, UK), using a λ = 633 nm laser at 0.4 mW with an exposure time of 25 s at a magnification of 50x (numerical aperture of 0.75). The microscope calibrated to polystyrene (reference material) and had an energy resolution of 2.5 cm-1. Spectra were averaged over 5 locations and considered between 500–3500 cm-1, enabling visualisation of the D and G scatter peaks. Post spectral processing included a 6th order polynomial fluorescence baseline correction and calculation of the I(D) / I(G) intensity ratio, this was completed using OriginPro 8.5.1 software (Origin Lab, USA).

ATR FT-IR.

GO dispersions were lyophilised, then a small amount of the resultant lyophilised powder was placed onto the ATR sample cell of a Tensor 27 Bruker spectrometer (Bruker UK Limited, UK) equipped with a 3000 Series TM High Stability Temperature Controller with RS232 Control (Specac, UK) set to 60 °C. Spectroscopic analysis was carried out between 750 cm-1 and 3500 cm-1. Before each measurement a background measurement was acquired. This was then subtracted from the subsequent measurement. Data processing was completed using OriginPro 8.5.1 software (Origin Lab, USA) which included normalisation of spectra to the C=C band at 1623 cm-1.

Animal handling procedures.

Eight-week-old C57BL/6 mice (18.1 ± 0.7 g) were obtained from Harlan (Oxfordshire, UK) and allowed to acclimatise within the facility for 1 week. Mice were maintained under a 12 h light/dark cycle at a steady temperature and humidity. Mice had access to food and water ad libitum for the duration of the investigation. Our experiments were conducted in accordance with the UK Home Office Code of Practice (1989) for the housing and care of animals in scientific procedures, in accordance with prior approval from the UK Home Office (Project licence number: P6C826A12). All procedures follow the ARRIVE guidelines. During the experiments mice received intranasal instillations with 30 μg of GO sheets (1 mg/mL in dextrose 5% aqueous solution) under light anaesthesia. The dose was selected as it was the minimum dose that was comfortably received by the animals and enabled confident detection of the Raman signature of GO within the tissues. Further handling details are included in each relevant subsection, below. As part of all experiments, we included age- matched negative controls treated with either dextrose 5% or DOTA[111In or 115In] alone, in order to verify and substantiate our findings. Throughout all our experiments, mice were weighed and their body condition was checked regularly.

247

Intranasal administration.

Mice were subjected to light anaesthesia via the inhalation of 2.5% isoflurane where oxygen was used as a carrier gas set at a flow rate of 2 L/min. Once sedated, mice were firmly held upright in a semi-fowlers position. As part of the single exposure experiment, using a calibrated Gilson pipette, a single bolus (30 μL intranasal dose of dispersion) was administered to the subject mice, half of the dose delivered in each nostril. For the multiple exposure experiments, the administrations were conducted as above, on a daily basis for 5 consecutive days, as per the OECD guidelines for sub-acute inhalation studies97.

Detection of GO in nasal cavities.

Mice were culled via CO2 asphyxiation and death was confirmed via the onset of rigor mortis. Mice nasal cavities were then carefully extracted and stored overnight in MilliQ water to disrupt and remove any blood cells that could obscure further analysis. Nasal cavities were then photographed at 20x magnification using a macro-lens (MPOW, UK). The extracted nasal cavities were then dried at 37 ⁰C, placed onto a glass slide and which was placed inside the sample compartment of a DXRi Raman microscope (ThermoScientific, USA). The nasal cavities were screened for the presence of GO as per its characteristic Raman signature. Raman spectra were recorded using laser excitation wavelength of 633 nm, a laser power of 0.4 mW, an exposure time of 25 seconds and were an average of 3 repeat measurements.

SPECT/CT whole body live imaging.

Mice were subjected to anaesthesia via the inhalation of 2.5% isoflurane where oxygen was used as a carrier gas set at a flow rate of 2 L/min. Each animal was then intranasally administered with an optimised dose of 30 μL of 1 mg/mL radio-labelled GO-DOTA[111In] (large, small or ultra-small sheet size) dispersion (half of the dose in each nostril) with an activity of approximately 5-6 MBq. At different time points post administration (1, 4 and 24 h), the brain biodistribution of each of the three materials was determined using a Nano-Scan® SPECT/CT scanner (Mediso, Hungary). All animals were anaesthetised by 4% isofluorane inhalation, prior and during the SPECT/CT imaging. SPECT images were collected over 20 projections in 40-60 min utilising a 4-head scanner with 1.4 mm pinhole collimators. X-ray CT scans were taken at the end of each SPECT acquisition with a semi-circular method with full scan, 480 projections, maximum FOV, 35 kV energy, 300 ms exposure time and 1-4 binning. Acquisitions were completed using the Nucline v2.01 (Build 020.0000) software (Mediso, Hungary), while reconstruction of all images and fusion of SPECT with CT images were achieved using the Interview™ FUSION bulletin software (Mediso,

248

Hungary). The images were further analysed with the VivoQuant 3.0 software (Boston, US), where the SPECT images were corrected for decay and for the minor differences in injected doses between animals.

Autoradiography of brains.

The brains of mice that were used for SPECT/CT imaging were extracted, washed in PBS and then fixed with 4% PFA in 1x PBS overnight. The next day the brains were cut into equal halves down the sagittal plane, dried briefly with paper tissue roll and placed with the flat- side down onto an autoradiography plate and left undisturbed for 7 days to develop. Autoradiography maps were then read using a Cyclone phosphor detector (Packard Biosciences, UK).

γ-scintigraphy of brain samples.

In order to allow for a more quantitative assessment of the tissue distribution of the GO- DOTA[111In] in the brain, a cut and count study was carried out. Mice were initially placed under light anaesthesia using 2.5% isoflurane where oxygen was used as a carrier gas set at a flow rate of 2 L/min. Then the mice were intranasally administered with an optimised dose of 30 μL of 1 mg/mL radio-labelled GO-DOTA[111In] dispersion (half of the dose in each nostril) with an activity of approximately 5-6 MBq. Mice were allowed to recover and were placed back into their assigned cages. After 24 h, mice were culled via cardiac perfusion using physiological saline supplemented with 0.002% porcine heparin sulphate (Sigma Aldrich, UK). The brains of mice were then extracted and micro-dissected to isolate the olfactory bulb, the cortex, striatum, hippocampus, mid brain, cerebellum, pons and medulla and any remaining part. Regular washing of blades was completed and careful attention was paid to not cross-contaminate adjacent substructures during dissection. Each of the stated substructures was placed into pre-weighed measurement tubes. All collected samples were counted on a γ-counter (Perkin Elmer, USA), along with dilutions of the injected dose with dead time limit below 60%. Following the measurements, all samples were dried and reweighed allowing the assessment of only dry weighed. The percentage of injected dose per gram dry tissue was calculated, using 3-4 different mice for each condition.

ICP-MS of brain samples.

In order to allow for a second quantitative assessment of the tissue distribution of us-GO- DOTA[115In] in the brain at 1 and 7 days post administration, brains were prepared for ICP- MS analysis. Mice were initially placed under light anaesthesia using 2.5% isoflurane where oxygen was used as a carrier gas set at a flow rate of 2 L/min. Mice were then intranasally

249 administered with an optimised dose of 30 μL of 1 mg/mL us-GO-DOTA[115In]. Mice were allowed to recover and were placed back into their assigned cages. Mice were culled via cardiac perfusion using physiological saline with 0.002% porcine heparin sulphate (Sigma Aldrich, UK) at 24 hours and 7 days post instillation. The brains of mice were extracted and micro-dissected as mentioned above. The organic components of the samples were oxidised by incubating the dried tissues with 1 mL of 30% hydrogen peroxide at 150 ⁰C, until the hydrogen peroxide evaporated. Following this, 1 mL of nitric acid was added and again left to evaporate. This hydrogen peroxide – nitric acid cycle was repeated 3 times. Hydrogen peroxide was then added to nitric acid at a 1:10 volume ratio resulted in the formation of a highly reactive oxidative solution, which helped remove any sample stuck to the bottom of the glass vessel. This was again left to evaporate at 150 ⁰C. Subsequently we performed 3 more successive hydrogen peroxide – nitric acid oxidative cycles. At this point the solution appeared clear, indicating the absence of organic matter. Finally, we left the solution to dry and dissolved the remaining inorganic materials in 5 mL of 2% nitric acid. Samples underwent analysis to quantitatively determine the amount of 115In present using an Agilent 7500cx ICP-MS (Agilent, UK). We included the original un-injected GO-DOTA[115In] control samples that had been processed in the same way. The percentage of injected dose per gram dry tissue was calculated, using 4 different mice for each condition. In all cases, we included dextrose-treated mice as negative controls and subtracted these from the test samples.

Immunohistochemistry of brain sections using fluorescent probes.

Mice were exposed to 30 μg of non-functionalised GO (1 μg/mL) as previously described. After 24 h, mice were anaesthetised with 2.5% isoflurane and subsequently and culled via cardiac perfusion with physiological saline with 0.002% porcine heparin sulphate (Sigma Aldrich, UK) followed by 4% paraformaldehyde (PFA). Whole brains were then extracted intact and submerged in 4% PFA and stored overnight. The brains were removed from the 4% PFA and stored at 4 ⁰C in 30% sucrose for at least 72 hours. Brains were next removed and immersed into the optimal cutting temperature (OCT) medium and snap-frozen in isopentane (Sigma, UK) pre-cooled to dry ice temperatures. The brains were subsequently subject to cryomicrotomy (Leica, UK) to obtain 16 µm sections which were placed onto glass slides and stored at -80⁰C until further processing. The tissue sections were further fixed using 300 µL ice cold methanol (Sigma-Aldrich, UK) for 10 min. The methanol was removed and the fixed tissues were incubated in blocking buffer (1% goat serum albumin (Sigma- Aldrich, UK) in 1 X PBS with 0.01% Triton 100x (Sigma-Aldrich, UK) for 1 h in a humidification chamber. Tissue sections were then placed in a washing buffer, consisting of

250

0.1% Triton X100 (Sigma, UK) in 1X PBS (Life technologies, UK), three times for 5 min. The washing buffer was removed and the tissues were separately incubated overnight at 4⁰C with the primary antibody rabbit raised anti-mouse IBA-1 (1:200) for microglia, mouse raised anti-mouse anti-NeuN (1:200) for neurones and chicken raised anti-mouse anti-GFAP (1:200) for astrocytes. The respective primary antibody solutions were then removed and the tissue gently rinsed with washing buffer three times for five min each. The washed sections were then incubated with the respective secondary antibody that included goat raised anti- chicken (1:200), goat raised anti-rabbit (1:200) and goat raised anti-mouse (1:200). In all staining experiments, secondary antibody controls as well as unstained controls were used to ensure the validity of our results.

Raman spectroscopy of brain samples.

In order to help validate the results of the γ-scintigraphy and ICP-MS based experiments, an equivalent, parallel study was carried out using non-functionalised GO material instead. Mice were initially placed under light anaesthesia using 2.5% isoflurane where oxygen was used as a carrier gas set at a flow rate of 2 L/min. Mice were then intranasally administered with 30 μL of 1 mg/mL of either l-GO, s-GO or us-GO. After the prescribed time period (1 day, 7 days or 1 month), mice were culled via cardiac perfusion using 30-40 mL of physiological saline with 0.002% porcine heparin sulphate (Sigma Aldrich, UK) under anaesthesia. The brains of the mice were then extracted and micro-dissected into the component substructures namely the olfactory bulb, the cortex, striatum, hippocampus, mid brain, cerebellum and pons and medulla. The isolated substructures were then mechanically homogenised onto glass slides, left to dry and then stored at -20 ⁰C to reduce the risk of microorganism growth which could hamper further analysis. Separate tools were used for the extraction of each substructure and all tools were washed following the microdissection of individual brains. This was completed to avoid the risk of cross contamination between sub-structures in each brain and between the brains of the different animals. A DXRxi confocal Raman microscope (ThermoScientific, USA) was then used to screen the samples for the presence of GO. GO was considered to be present in the sample following the successful acquisition of regions that screened Raman positive for GO. For each micro- dissected brain region, a total of 5 regions were scanned. For each condition an n = 3 mice were used. A total of 15 spectra were therefore analysed per brain substructure per time point. GO spectra were scrutinised to assess the ratio between the intensity of the D peak and the intensity of G peak, I(D)/ I(G), over time. In all cases, mice treated with 30 μL of dextrose 5 % were used as negative controls and were screened in the same way. This was

251 completed to ensure the quality of our experiments and the lack of cross-contamination between samples.

Statistical analysis. All animals experiments were completed with an n = 3-4 repeats. The statistical differences between multiple groups were assessed using the Kruskal-Wallis test followed by Dunn’s post-hoc test. Statistical differences were considered using a p value of < 0.05 and performed using GraphPad Prism.

6.7. Acknowledgments.

This work was partially supported by the EPSRC NowNano DTC program (grants EP/K016946/1 and EP/M010619/1) , and the EU 7th RTD Framework Programme, Graphene Flagship project (FP7-ICT-2013-FET-F-604391). We would like to acknowledge the staff in the Faculty of Biology, Medicine and Health EM Facility, particularly Dr Aleksandr Mironov and Ms Samantha Forbes for their expert advice and assistance, and the Wellcome Trust for equipment grant support to the EM Facility. The University of Manchester Bioimaging Facility microscopes used in this study were purchased with grants from the BBSRC, Wellcome Trust and the University of Manchester Strategic Fund. We wish to thank Mr R. Meadows from the Bio-imaging Facility. We are also grateful to Mr P. Walker, from the Histology Facility, University of Manchester for his expert advice and kind assistance in the tissue histology. We thank Dr N. Hodson from the Bio-AFM Facility (Centre for Tissue Injury and Repair at the University of Manchester) for assistance and advice regarding the AFM instrumentation. X-ray photoelectron spectroscopy was performed at the National EPSRC XPS User’s Service (NEXUS) at Newcastle University, an EPSRC Mid-Range Facility.

252

6.8. Supplementary Information.

Parameter Technique l-GO s-GO us-GO Optical 1 μm – 32 μm - - Microscopy Lateral Dimension AFM 5 μm – 20 μm 30 nm – 700 nm 30 nm – 300 nm

TEM 0.5 μm – 15 μm 50 nm – 3 μm 20 nm – 200 nm

Thickness AFM 1 nm – 5 nm 1 nm – 7 nm 1 nm – 4 nm Degree of defects Raman 1.30 0.09 1.35 0.03 1.34 0.03 I (D)/I (G) Surface charge ζ-potential -61.7 1.8 mV -55.0 0.8 mV -34.8 0.1 mV Functionalisation TGA 37% 39% 37% Degree Chemical C: 68.3%, O: 31.5%, C: 67.8%, O: 31.8%, C: 68.4%, O: 31.0%, XPS composition N: 0.3% N: 0.4%, Na: 0.1% N: 0.4%, Na: 0.3%

Purity (O:C ratio) XPS 99.7% (0.5) 99.5% (0.5) 99.4% (0.5) π-π*, O=C-O, C=O, 1.3%, 16.8%, 33.1%, 1.3%, 15.1%, 31.2%, 1.7%, 15.4%, 28.9%, XPS C-O, C-C & C=C 10.1%, 38.7% 12.7%, 39.8% 9.6%, 44.4%

Filipe Rodrigues contributed this table.

253

GO-PEG -DOTA GO NH -PEG -DOTA 4 2 4 (GO-DOTA)

Figure 6.S1 Schematic illustrating the epoxide ring-opening reaction mechanism which we used to functionalise GO sheets (l-GO, s-GO or us-GO) mechanism with DOTA.

For the sake of clarity, we only show one epoxide group being functionalised by PEG4-DOTA. Isabella Vacchi contributed this Figure and performed the reaction.

254

l-GO s-GO us-GO 20 30 20 Non- Functionalised

10 15 10

0 0 0

0 10 20 30 40 0 1000 2000 0 100 200 300 400

% Distribution, % 100 = n % Distribution, % 300 = n Lateral Dimensions (m) Distribution, % 100 = n Lateral Dimensions (nm) Lateral Dimensions (nm) 20 20 20 TEM DOTA- Functionalised

10 10 10 Lateral Dimension Lateral

0 0 0

0 1 2 3 4 5 6 0 500 1000 1500 0 200 400 600

% Distribution, % 119 = n % Distribution, % 116 = n Lateral Dimension (m) Lateral Dimension (nm) Distribution, % 100 = n Lateral Dimension (nm) 40 30 40 Non- Functionalised

20 15 20

0 0 0

0 20 40 0 100 200 300 400 0 100 200 300 % Distribution, % 12 = n

Lateral Dimension (m) Distribution, % 750 = n Lateral Dimension (nm) Lateral Dimension (nm) % Distribution, % 1635 = n 50 40 60 DOTA- 40 48 30 Functionalised

Lateral Dimension Lateral 30 36 20 20 24 10 10 12

0 0 0

0 1 2 3 4 0 200 400 600 800 0 100 200 300 400 500

% Distribution, % 421 = n % Distribution, % 166 = n

Lateral Dimension (µm) Distribution, % 618 = n Lateral Dimension (nm) Lateral Dimension (nm)

AFM 70 50 40 Non- Functionalised

35 25 20

0 0 0

0 2 4 6 8 0 4 8 12 0 1 2 3 % Distribution, % 12 = n

% Distribution, % 738 = n Thickness (nm) Thickness (nm) Distribution, % 262 = n Thickness (nm)

50 60 80 DOTA- Thickness 40 48 60 Functionalised 30 36 40 20 24 20 10 12

0 0 0

0 10 20 30 40 50 60 0 5 10 15 20 25 30 35 0 10 20 30 40 50

% Distribution, % 618 = n % Distribution, % 421 = n % Distribution, % 166 = n Thickness (nm) Thickness (nm) Thickness (nm)

Figure 6.S2 Dimensional analysis of starting and DOTA functionalised l-GO, s-GO and us-GO.

The histograms correspond to the TEM (for lateral dimensions) and AFM data (for lateral dimensions and thickness) given in Figure 6.1.

255

s-GO l-GO us-GO

l-GO-DOTA s-GO-DOTA us-GO DOTA Raman

1000 2000 3000 (a.u.) Intensity Raman 1000 2000 3000 Raman Intensity (a.u.) Intensity Raman 1000 2000 3000

-1 (a.u.) Intensity Raman Raman Shift (cm-1) Raman Shift (cm ) Raman Shift (cm-1)

l-GO s-GO us-GO

FTIR l-GO-DOTA

- s-GO-DOTA us-GO DOTA

ATR

Absorbance (a.u.) Absorbance

Absorbance (a.u.) Absorbance Absorbance (a.u.) Absorbance 4000 3000 2000 1000 4000 3000 2000 1000 4000 3000 2000 1000 Wavenumber (cm-1) -1 Wavenumber (cm-1) Wavenumber (cm )

Figure 6.S3 Spectroscopic features of l-GO, s-GO and us-GO materials and their respective DOTA functionalised derivatives as analysed using Raman spectroscopy and ATR FT-IR.

256

l-GO-DOTA[111In] s-GO-DOTA[111In] us-GO-DOTA[111In] DOTA[111In]

30 minutes30 3 hours3

Figure 6.S4 SPECT/CT-based qualitative biodistribution at early time points (30 mins and 3 hours post administration).

Dr Dhifaf Jasim performed the SPECT / CT.

l-GO s-GO us-GO Dex 5%

B l-GO s-GO us-GO Dex 5%

n.d.

1000 2000 3000

Raman Intensity (a.u.) Intensity Raman 1000 2000 3000

Raman Intensity (a.u.) Intensity Raman 1000 2000 3000 Raman Intensity (a.u.) Intensity Raman -1 -1 Raman Shift (cm ) Raman Shift (cm-1) Raman Shift (cm-1) Raman Shift (cm )

Figure 6.S5 The presence of GO in the extracted nasal cavities of mice after 24 hours following treatment with unlabelled l-GO, s-GO and us-GO.

(A) Photographs of extracted nasal cavities were taken to highlight the presence of entrapped dark material. (B) Raman spectroscopy was employed to determine the graphitic nature of the material.

257

Raman Intensity (a.u.) RamanIntensity

1000 2000 3000 Raman Shift (cm-1)

Figure 6.S6 Raman spectra used as reference for Raman-based correlative mapping of brain sections as shown in Figure 6.4.

258

Olfactory bulb Cerebellum

500 μm 500 μm

(Microglia) (Microglia) (Microglia) (Microglia)

– –

IHC IHC IHC IHC

500 μm 500 μm

(Astrocytes) (Astrocytes)

–

IHC IHC IHC IHC

500 μm 500 μm

(Neurons) (Neurons)

–

IHC IHC IHC IHC

Figure 6.S7 Bright field images showing regions of interest where confocal Raman maps were acquired.

Images were taken directly from the Thermo Scientific Raman imaging software. Dotted rectangles indicate the regions that underwent Raman mapping analysis and correspond to Figure 6.5. Labels on the left of each image are given for clarity to indicate the image in Figure 4 that each respective image corresponds to.

259

0.04

Day 1 (us-GO-DOTA[115In]) Day 7 (us-GO-DOTA[115In]) Day 1 (DOTA[115In]) Day 7 (DOTA[115In])

0.02

Percentageof Administereddose (%) 0.00 6 Cortex Striatum Midbrain Cerebellum Remaining Olfactory Bulb Hippocampus Pons and Medulla

Figure 6.S8 3Changes in the brain biodistribution profile of us-GO at day 1 and 7 following a single intranasal administration.

The brain biodistribution profile of us-GO-DOTA[115In] at days 1 and 7 post administration was determined following a single intranasal administration of us-GO. Each micro-dissected substructure was analysed with ICP-MS for the presence of 115In. The data has been expressed as percentage of the administered0 dose (%) in each brain anatomical region and subtracted from negative control values. Statistical significance was assessed using a Kruskal-Wallis test with post hoc Dunn’s multiple Cortex comparisons test. *: p < 0.05Striatum, n = 4 mouseMidbrain were used per group. Cerebellum Remaining Olfactory Bulb Hippocampus Pons and Medulla

260

A B Single Exposure Multiple Exposure §§ §§

§ ) 1.50 2 § § * 20 μm 0.75

1.50

countsmm / (

0.00 0.75

OB (Day 1)OB (DayOB 7) (Day 28)CB (Day 1)CB (DayCB 7) (Day 28) tissue section section tissue GO Raman positive regions per positive Raman GO regions

Figure 6.S9 Presence of GO in the brain decreases over 28 days following intranasal 0.00 instillation.

(A)The presence of GO in the olfactory bulb (OB) and cerebellum (CB) was quantified in brain sections by determining the number of areas which were positive for the characteristic Raman signature (shown in Figure S6), following single (green) or multiple (dark green) intranasal administrations of us-GO. At each time point, 2 sections from each mice (n = 3 mice) were screen for us-GO. (B) The typical bright field appearance of a GO active site is also shown. Statistical differences between single and multiple exposures, in terms of the number of GO Raman positive regions per mm2, were assessed using the Mann-Whitney U test, *: p < 0.05; while statistical differences between the number of GO Raman positive regions per mm2 at different time points following single or multiple exposures were assessed using the Kruskal-Wallis test with post hoc Dunn’s multiple comparisons test, §: p < 0.05, §§: p < 0.01.

261

A Olfactory Bulb Cerebellum )

) 15 15

a.u.

3 3 10

10 x10

x10

( (

5 5 Day 7 Day 28 0 0 1000 1500 2000 1000 1500 2000 -1 -1

Raman Intensity Raman Raman Shift (cm )

B Olfactory Bulb Cerebellum

2.5 2.5 15 2.0 **** 2.0 **

1.5 1.5 10 I(D) / I(G) / I(D)

I(D) / I(G) / I(D) 1.0 1.0 5 0.5 0.5 7 28 7 28 Time (Days) Time (Days) 0 1000 1500 2000 Figure 6.S10 Degradation of us-GO 28 days following multiple intranasal administrations.

The olfactory bulb (left) and cerebellum (right) of mice treated with us-GO were isolated and physically homogenised at days 7 and 28. The tissues were then screened for the presence of GO as per the materials characteristic Raman spectra. The average Raman signal intensity and the box plots of the I(D) / I(G) are given. At each time point n = 5 spectra were collected from each of the 3 treated mice. Statistical significance was assessed using a Kruskal-Wallis test with post hoc Dunn’s multiple comparisons test. *: p < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001.

262

25 Single Exposure (us-GO) Single exposure (Dex 5%) Multiple Exposure (us-GO) Multiple exposure (Dex 5%)

20 Single administration Weight Weight (g) Mouse of

15 Multiple administrations 0 10 20 30 Time (Days)

Figure 6.S11 Growth curve of C57BL/6 mice administered instranasally with us-GO on a single occasion or on multiple occasions.

Blue arrows indicate times of administrations. All points on the curve represent averages of n = 3 mice.

263

6.9. References

1. D. Jariwala, V. K. Sangwan, L. J. Lauhon, T. J. Marks and M. C. Hersam, Chem. Soc. Rev., 2013, 42, 2824-2860. Carbon nanomaterials for electronics, optoelectronics, photovoltaics, and sensing. 2. A. C. Ferrari, F. Bonaccorso, V. Fal'ko, K. S. Novoselov, S. Roche, P. Boggild, S. Borini, F. H. L. Koppens, V. Palermo, N. Pugno, J. A. Garrido, R. Sordan, A. Bianco, L. Ballerini, M. Prato, E. Lidorikis, J. Kivioja, C. Marinelli, T. Ryhanen, A. Morpurgo, J. N. Coleman, V. Nicolosi, L. Colombo, A. Fert, M. Garcia-Hernandez, A. Bachtold, G. F. Schneider, F. Guinea, C. Dekker, M. Barbone, Z. Sun, C. Galiotis, A. N. Grigorenko, G. Konstantatos, A. Kis, M. Katsnelson, L. Vandersypen, A. Loiseau, V. Morandi, D. Neumaier, E. Treossi, V. Pellegrini, M. Polini, A. Tredicucci, G. M. Williams, B. Hee Hong, J.-H. Ahn, J. Min Kim, H. Zirath, B. J. van Wees, H. van der Zant, L. Occhipinti, A. Di Matteo, I. A. Kinloch, T. Seyller, E. Quesnel, X. Feng, K. Teo, N. Rupesinghe, P. Hakonen, S. R. T. Neil, Q. Tannock, T. Lofwander and J. Kinaret, Nanoscale, 2015, 7, 4598-4810. Science and technology roadmap for graphene, related two-dimensional crystals, and hybrid systems. 3. K. S. Novoselov, V. I. Falko, L. Colombo, P. R. Gellert, M. G. Schwab and K. Kim, Nature, 2012, 490, 192-200. A roadmap for graphene. 4. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666-669. Electric field effect in atomically thin carbon films. 5. X. Huang, Z. Yin, S. Wu, X. Qi, Q. He, Q. Zhang, Q. Yan, F. Boey and H. Zhang, Small, 2011, 7, 1876-1902. Graphene-based materials: synthesis, characterization, properties, and applications. 6. D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, Chem. Soc. Rev., 2010, 39, 228-240. The chemistry of graphene oxide. 7. Y. Su, V. G. Kravets, S. L. Wong, J. Waters, A. K. Geim and R. R. Nair, Nat. Commun., 2014, 5, 4843. Impermeable barrier films and protective coatings based on reduced graphene oxide. 8. J. Abraham, K. S. Vasu, C. D. Williams, K. Gopinadhan, Y. Su, C. T. Cherian, J. Dix, E. Prestat, S. J. Haigh, I. V. Grigorieva, P. Carbone, A. K. Geim and R. R. Nair, Nat. Nanotechnol., 2017, 12, 546-550. Tunable sieving of ions using graphene oxide membranes. 9. F. Li, X. Jiang, J. Zhao and S. Zhang, Nano Energy, 2015, 16, 488-515. Graphene oxide: A promising nanomaterial for energy and environmental applications.

264

10. D. Bitounis, H. Ali-Boucetta, B. H. Hong, D.-H. Min and K. Kostarelos, Adv. Mater., 2013, 25, 2258-2268. Prospects and challenges of graphene in biomedical applications. 11. S. Gurunathan and J.-H. Kim, Int. J. Nanomed., 2016, 11, 1927-1945. Synthesis, toxicity, biocompatibility, and biomedical applications of graphene and graphene- related materials. 12. V. C. Sanchez, A. Jachak, R. H. Hurt and A. B. Kane, Chem. Res. Toxicol., 2012, 25, 15-34. Biological interactions of graphene-family nanomaterials: an interdisciplinary review. 13. M. J. Ellenbecker and C. S.-J. Tsai, in Exposure Assessment and Safety Considerations for Working with Engineered Nanoparticles, John Wiley & Sons, Inc, Hoboken, 2015, Ch. 4, pp. 39-50. 14. ICRP, Ann ICRP., 1994, 24, 1-482. Human respiratory tract model for radiological protection. 15. L. Illum, J. Pharm. Pharmacol., 2004, 56, 3-17. Is nose-to-brain transport of drugs in man a reality? 16. W. G. Kreyling, M. Semmler-Behnke and W. Moller, J. Aerosol Med., 2006, 19, 74- 83. Ultrafine particle-lung interactions: does size matter? 17. W. G. Kreyling, M. Semmler-Behnke and W. Möller, J. Nanopart. Res., 2006, 8, 543- 562. Health implications of nanoparticles. 18. G. Oberdörster, A. Elder and A. Rinderknecht, J. Nanosci. Nanotechnol., 2009, 9, 4996-5007. Nanoparticles and the brain: cause for concern? 19. A. J. De Lorenzo, J Biophys Biochem Cytol, 1957, 3, 839-850. Electron microscopic observations of the olfactory mucosa and olfactory nerve. 20. E. S. Patchin, D. S. Anderson, R. M. Silva, D. L. Uyeminami, G. M. Scott, T. Guo, L. S. Van Winkle and K. E. Pinkerton, Environ. Health Perspect., 2016, 124, 1870-1875. Size-dependent deposition, translocation, and microglial activation of inhaled silver nanoparticles in the rodent nose and brain. 21. B. Wang, W. Y. Feng, M. Wang, J. W. Shi, F. Zhang, H. Ouyang, Y. L. Zhao, Z. F. Chai, Y. Y. Huang, Y. N. Xie, H. F. Wang and J. Wang, Biol. Trace Elem. Res., 2007, 118, 233-243. Transport of intranasally instilled fine Fe2O3 particles into the brain: micro-distribution, chemical states, and histopathological observation. 22. G. Oberdorster, Z. Sharp, V. Atudorei, A. Elder, R. Gelein, W. Kreyling and C. Cox, Inhalation Toxicol., 2004, 16, 437-445. Translocation of inhaled ultrafine particles to the brain.

265

23. J. Wang, C. Chen, Y. Liu, F. Jiao, W. Li, F. Lao, Y. Li, B. Li, C. Ge, G. Zhou, Y. Gao, Y. Zhao and Z. Chai, Toxicol. Lett. 2008, 183, 72-80. Potential neurological lesion after nasal instillation of TiO(2) nanoparticles in the anatase and rutile crystal phases. 24. A. Elder, R. Gelein, V. Silva, T. Feikert, L. Opanashuk, J. Carter, R. Potter, A. Maynard, Y. Ito, J. Finkelstein and G. Oberdorster, Environ. Health Perspect., 2006, 114, 1172-1178. Translocation of inhaled ultrafine manganese oxide particles to the central nervous system. 25. X. Zhuang, X. Xiang, W. Grizzle, D. Sun, S. Zhang, R. C. Axtell, S. Ju, J. Mu, L. Zhang, L. Steinman, D. Miller and H. G. Zhang, Mol. Ther. 2011, 19, 1769-1779. Treatment of brain inflammatory diseases by delivering exosome encapsulated anti- inflammatory drugs from the nasal region to the brain. 26. B. A. Maher, I. A. M. Ahmed, V. Karloukovski, D. A. MacLaren, P. G. Foulds, D. Allsop, D. M. A. Mann, R. Torres-Jardón and L. Calderon-Garciduenas, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 10797-10801. Magnetite pollution nanoparticles in the human brain. 27. R. G. Thorne, G. J. Pronk, V. Padmanabhan and W. H. Frey, 2nd, Neuroscience, 2004, 127, 481-496. Delivery of insulin-like growth factor-I to the rat brain and spinal cord along olfactory and trigeminal pathways following intranasal administration 28. D. D. Hunter and R. D. Dey, Neuroscience, 1998, 83, 591-599. Identification and neuropeptide content of trigeminal neurons innervating the rat nasal epithelium. 29. W. H. Frey, J. Liu, X. Chen, R. G. Thorne, J. R. Fawcett, T. A. Ala and Y.-E. Rahman, Drug delivery, 1997, 4, 87-92. Delivery of 125I-NGF to the brain via the olfactory route. 30. G. Shepherd in Neurobiology, Oxford University Press, New York, 3nd edn, 1994, Ch. 9, pp. 221-224. 31. L. Illum, Eur. J. Pharm. Sci., 2000, 11, 1-18. Transport of drugs from the nasal cavity to the central nervous system. 32. W. M. Pardridge, Drug Discovery Today, 2007, 12, 54-61. Blood-brain barrier delivery. 33. J. A. Loureiro, B. Gomes, M. A. Coelho, M. do Carmo Pereira and S. Rocha, Nanomedicine (Lond.), 2014, 9, 709-722. Targeting nanoparticles across the blood– brain barrier with monoclonal antibodies. 34. A. Schinwald, F. A. Murphy, A. Jones, W. MacNee and K. Donaldson, ACS Nano, 2012, 6, 736-746. Graphene-based nanoplatelets: a new risk to the respiratory system as a consequence of their unusual aerodynamic properties. 35. C. Bussy, D. Jasim, N. Lozano, D. Terry and K. Kostarelos, Nanoscale, 2015, 7, 6432-6435. The current graphene safety landscape – a literature mining exercise.

266

36. C. M. Girish, A. Sasidharan, G. S. Gowd, S. Nair and M. Koyakutty, Adv. Healthcare Mater., 2013, 2, 1489-1500. Confocal Raman imaging study showing macrophage mediated biodegradation of graphene in vivo. 37. A. Nunes, C. Bussy, L. Gherardini, M. Meneghetti, M. A. Herrero, A. Bianco, M. Prato, T. Pizzorusso, K. T. Al-Jamal and K. Kostarelos, Nanomedicine (Lond.), 2012, 7, 1485-1494. In vivo degradation of functionalized carbon nanotubes after stereotactic administration in the brain cortex. 38. A. A. Shvedova, A. A. Kapralov, W. H. Feng, E. R. Kisin, A. R. Murray, R. R. Mercer, C. M. St Croix, M. A. Lang, S. C. Watkins, N. V. Konduru, B. L. Allen, J. Conroy, G. P. Kotchey, B. M. Mohamed, A. D. Meade, Y. Volkov, A. Star, B. Fadeel and V. E. Kagan, PloS One, 2012, 7, e30923. Impaired clearance and enhanced pulmonary inflammatory/fibrotic response to carbon nanotubes in myeloperoxidase-deficient mice. 39. D. Elgrabli, W. Dachraoui, C. Ménard-Moyon, X. J. Liu, D. Bégin, S. Bégin-Colin, A. Bianco, F. Gazeau and D. Alloyeau, ACS Nano, 2015, 9, 10113-10124. Carbon nanotube degradation in macrophages: live nanoscale monitoring and understanding of biological pathway. 40. C. Bussy, C. Hadad, M. Prato, A. Bianco and K. Kostarelos, Nanoscale, 2016, 8, 590-601. Intracellular degradation of chemically functionalized carbon nanotubes using a long-term primary microglial culture model. 41. A. E. Goode, D. A. Gonzalez Carter, M. Motskin, I. S. Pienaar, S. Chen, S. Hu, P. Ruenraroengsak, M. P. Ryan, M. S. P. Shaffer, D. T. Dexter and A. E. Porter, Biomaterials, 2015, 70, 57-70. High resolution and dynamic imaging of biopersistence and bioreactivity of etra and intracellular MWNTs exposed to microglial cells. 42. J. Russier, L. Oudjedi, M. Piponnier, C. Bussy, M. Prato, K. Kostarelos, B. Lounis, A. Bianco and L. Cognet, Nanoscale, 2017, 9, 4642-4645. Direct visualization of carbon nanotube degradation in primary cells by photothermal imaging. 43. S. P. Mukherjee, N. Lozano, M. Kucki, A. E. Del Rio-Castillo, L. Newman, E. Vázquez, K. Kostarelos, P. Wick and B. Fadeel, PloS One, 2016, 11, e0166816. Detection of endotoxin contamination of graphene based materials using the TNF-α expression test and guidelines for endotoxin-free graphene oxide production. 44. R. Rauti, N. Lozano, V. León, D. Scaini, M. Musto, I. Rago, F. P. Ulloa Severino, A. Fabbro, L. Casalis, E. Vázquez, K. Kostarelos, M. Prato and L. Ballerini, ACS Nano, 2016, 10, 4459-4471. Graphene oxide nanosheets reshape synaptic function in cultured brain networks.

267

45. D. A. Jasim, C. Menard-Moyon, D. Begin, A. Bianco and K. Kostarelos, Chem. Sci. 2015, 6, 3952-3964. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. 46. A. Peters and C. A. Pope, 3rd, Lancet, 2002, 360, 1184-1185. Cardiopulmonary mortality and air pollution. 47. D. Mittal, A. Ali, S. Md, S. Baboota, J. K. Sahni and J. Ali, Drug delivery, 2014, 21, 75-86. Insights into direct nose to brain delivery: Current status and future perspective. 48. J. Wu, T. Ding and J. Sun, Neurotoxicology, 2013, 34, 243-253. Neurotoxic potential of iron oxide nanoparticles in the rat brain striatum and hippocampus. 49. I. A. Vacchi, C. Spinato, J. Raya, A. Bianco and C. Menard-Moyon, Nanoscale, 2016, 8, 13714-13721. Chemical reactivity of graphene oxide towards amines elucidated by solid-state NMR. 50. D. A. Jasim, C. Menard-Moyon, D. Begin, A. Bianco and K. Kostarelos, Chem. Sci., 2015. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. 51. Y. Li, L. Feng, X. Shi, X. Wang, Y. Yang, K. Yang, T. Liu, G. Yang and Z. Liu, Small, 2013. Surface coating-dependent cytotoxicity and degradation of graphene derivatives: towards the design of non-toxic, degradable nano-graphene. 52. J. L. Li, K. N. Kudin, M. J. McAllister, R. K. Prud'homme, I. A. Aksay and R. Car, Phys. Rev. Lett., 2006, 96, 176101. Oxygen-driven unzipping of graphitic materials. 53. T. Sun and S. Fabris, Nano Lett., 2012, 12, 17-21. Mechanisms for oxidative unzipping and cutting of graphene. 54. K. Yang, H. Gong, X. Shi, J. Wan, Y. Zhang and Z. Liu, Biomaterials, 2013, 34, 2787- 2795. In vivo biodistribution and toxicology of functionalized nano-graphene oxide in mice after oral and intraperitoneal administration. 55. S. Zhang, K. Yang, L. Feng and Z. Liu, Carbon, 2011, 49, 4040-4049. In vitro and in vivo behaviors of dextran functionalized graphene. 56. A. C. Ferrari and J. Robertson, Phys. Rev. B, 2000, 61, 14095-14107. Interpretation of Raman spectra of disordered and amorphous carbon. 57. L. Du, S. Wu, Y. Li, X. Zhao, X. Ju and Y. Wang, Bio-Med. Mater. Eng., 2014, 24, 2135-2141. Cytotoxicity of PEGylated graphene oxide on lymphoma cells. 58. T. Banerjee, S. Mitra, A. Kumar Singh, R. Kumar Sharma and A. Maitra, Int. J. Pharm., 2002, 243, 93-105. Preparation, characterization and biodistribution of ultrafine chitosan nanoparticles. 59. E. D. Pressly, R. Rossin, A. Hagooly, K. Fukukawa, B. W. Messmore, M. J. Welch, K. L. Wooley, M. S. Lamm, R. A. Hule, D. J. Pochan and C. J. Hawker,

268

Biomacromolecules, 2007, 8, 3126-3134. Structural effects on the biodistribution and positron emission tomography (PET) imaging of well-defined (64)Cu-labeled nanoparticles comprised of amphiphilic block graft copolymers. 60. K. Yang, J. Wan, S. Zhang, Y. Zhang, S.-T. Lee and Z. Liu, ACS Nano, 2011, 5, 516- 522. In vivo pharmacokinetics, long-term biodistribution, and toxicology of PEGylated graphene in mice. 61. B. Cornelissen, S. Able, V. Kersemans, P. A. Waghorn, S. Myhra, K. Jurkshat, A. Crossley and K. A. Vallis, Biomaterials, 2013, 34, 1146-1154. Nanographene oxide- based radioimmunoconstructs for in vivo targeting and SPECT imaging of HER2- positive tumors. 62. P. Wick, A. E. Louw-Gaume, M. Kucki, H. F. Krug, K. Kostarelos, B. Fadeel, K. A. Dawson, A. Salvati, E. Vázquez, L. Ballerini, M. Tretiach, F. Benfenati, E. Flahaut, L. Gauthier, M. Prato and A. Bianco, Angew. Chem., Int. Ed., 2014, 53, 7714-7718. Classification framework for graphene-based materials. 63. C. Bussy, D. A. Jasim, N. Lozano, D. Terry and K. Kostarelos, Nanoscale, 2015, 7, 6432-6435. The current graphene safety landscape--a literature mining exercise. 64. L. E. Hopkins, E. S. Patchin, P.-L. Chiu, C. Brandenberger, S. Smiley-Jewell and K. E. Pinkerton, Nanotoxicology, 2014, 8, 885-893. Nose-to-brain transport of aerosolised quantum dots following acute exposure. 65. K. Kristensson and Y. Olsson, Acta Neuropathol., 1971, 19, 145-154. Uptake of exogenous proteins in mouse olfactory cells. 66. Y. Wang, R. Aun and F. L. Tse, Biopharm. Drug Dispos., 1998, 19, 571-575. Brain uptake of dihydroergotamine after intravenous and nasal administration in the rat. 67. H. S. Chow, Z. Chen and G. T. Matsuura, J. Pharm. Sci., 1999, 88, 754-758. Direct transport of cocaine from the nasal cavity to the brain following intranasal cocaine administration in rats. 68. J. J. Lochhead, D. J. Wolak, M. E. Pizzo and R. G. Thorne, J. Cereb. Blood Flow Metab., 2015, 35, 371-381. Rapid transport within cerebral perivascular spaces underlies widespread tracer distribution in the brain after intranasal administration. 69. B. R. Stevenson, J. M. Anderson and S. Bullivant, Mol. Cell. Biochem., 1988, 83, 129-145. The epithelial tight junction: structure, function and preliminary biochemical characterization. 70. H. R. Costantino, L. Illum, G. Brandt, P. H. Johnson and S. C. Quay, Int. J. Pharm., 2007, 337, 1-24. Intranasal delivery: physicochemical and therapeutic aspects. 71. M. van Woensel, N. Wauthoz, R. Rosière, K. Amighi, V. Mathieu, F. Lefranc, S. W. van Gool and S. de Vleeschouwer, Cancers, 2013, 5, 1020-1048. Formulations for

269

intranasal delivery of pharmacological agents to combat brain disease: A new opportunity to tackle GBM? 72. M. Caggiano, J. S. Kauer and D. D. Hunter, Neuron, 1994, 13, 339-352. Globose basal cells are neuronal progenitors in the olfactory epithelium: a lineage analysis using a replication-incompetent retrovirus. 73. C. Geraci, D. Heidel, C. Sayes, L. Hodson, P. Schulte, A. Eastlake and S. Brenner, J. Nanopart. Res., 2015, 17, 366. Perspectives on the design of safer nanomaterials and manufacturing processes. 74. M. S. Ali and J. P. Pearson, Laryngoscope, 2007, 117, 932-938. Upper airway mucin gene expression: a review. 75. S. K. Lai, Y.-Y. Wang and J. Hanes, Adv. Drug Delivery Rev., 2009, 61, 158-171. Mucus-penetrating nanoparticles for drug and gene delivery to mucosal tissues. 76. P. Hadaczek, Y. Yamashita, H. Mirek, L. Tamas, M. C. Bohn, C. Noble, J. W. Park and K. Bankiewicz, Mol. Ther., 2006, 14, 69-78. The “perivascular pump” driven by arterial pulsation is a powerful mechanism for the distribution of therapeutic molecules within the brain. 77. J. J. Iliff, M. Wang, Y. Liao, B. A. Plogg, W. Peng, G. A. Gundersen, H. Benveniste, G. E. Vates, R. Deane, S. A. Goldman, E. A. Nagelhus and M. Nedergaard, Sci. Transl. Med., 2012, 4, 147ra111. A paravascular pathway facilitates CSF flow through the brain parenchyma and the clearance of interstitial solutes, including amyloid β. 78. L. Yan, H. Wang, Y. Jiang, J. Liu, Z. Wang, Y. Yang, S. Huang and Y. Huang, Macromol. Res., 2013, 21, 435-441. Cell-penetrating peptide-modified PLGA nanoparticles for enhanced nose-to-brain macromolecular delivery. 79. D. A. Jasim, S. Murphy, L. Newman, A. Mironov, E. Prestat, J. McCaffrey, C. Ménard-Moyon, A. F. Rodrigues, A. Bianco, S. Haigh, R. Lennon and K. Kostarelos, ACS Nano, 2016, 10, 10753-10767. The effects of extensive glomerular filtration of thin graphene oxide sheets on kidney physiology. 80. X. Zhang, J. Yin, C. Peng, W. Hu, Z. Zhu, W. Li, C. Fan and Q. Huang, Carbon, 2011, 49, 986-995. A comparative study of cellular uptake and cytotoxicity of multiwalled carbon nanotubes, graphene oxide, and nanodiamond. 81. A. Sasidharan, S. Swaroop, C. K. Koduri, C. M. Girish, P. Chandran, L. S. Panchakarla, V. H. Somasundaram, G. S. Gowd, S. Nair and M. Koyakutty, Carbon, 2015, 95, 511-524. Comparative in vivo toxicity, organ biodistribution and immune response of pristine, carboxylated and PEGylated few-layer graphene sheets in Swiss albino mice: A three month study.

270

82. E. Hutter, S. Boridy, S. Labrecque, M. Lalancette-Hebert, J. Kriz, F. M. Winnik and D. Maysinger, ACS Nano, 2010, 4, 2595-2606. Microglial response to gold nanoparticles. 83. Y. Wang, B. Wang, M.-T. Zhu, M. Li, H.-J. Wang, M. Wang, H. Ouyang, Z.-F. Chai, W.-Y. Feng and Y.-L. Zhao, Toxicol. Lett., 2011, 205, 26-37. 84. C. Bussy, K. T. Al-Jamal, J. Boczkowski, S. Lanone, M. Prato, A. Bianco and K. Kostarelos, ACS Nano, 2015, 9, 7815-7830. 85. R. Nicolete, D. F. dos Santos and L. H. Faccioli, Int. Immunopharmacol., 2011, 11, 1557-1563. The uptake of PLGA micro or nanoparticles by macrophages provokes distinct in vitro inflammatory response. 86. B. Bedussi, M. G. J. T. B. van Lier, J. W. Bartstra, J. de Vos, M. Siebes, E. VanBavel and E. N. T. P. Bakker, Fluids Barriers CNS, 2015, 12, 23. Clearance from the mouse brain by convection of interstitial fluid towards the ventricular system. 87. L. Xie, H. Kang, Q. Xu, M. J. Chen, Y. Liao, M. Thiyagarajan, J. O'Donnell, D. J. Christensen, C. Nicholson, J. J. Iliff, T. Takano, R. Deane and M. Nedergaard, Science, 2013, 342, 373-377. Sleep drives metabolite clearance from the adult brain. 88. Y. Ding, R. Tian, Z. Yang, J. Chen and N. Lu, J. Mater. Sci.: Mater. Med., 2017, 28, 7. NADPH oxidase-dependent degradation of single-walled carbon nanotubes in macrophages. 89. K. Bhattacharya, S. P. Mukherjee, A. Gallud, S. C. Burkert, S. Bistarelli, S. Bellucci, M. Bottini, A. Star and B. Fadeel, Nanomedicine: Nanotechnology, Biology and Medicine, 2016, 12, 333-351. Biological interactions of carbon-based nanomaterials: From coronation to degradation. 90. M. Chen, X. Qin and G. Zeng, Trends Biotechnol., 2017, 35, 836-846. Biodegradation of carbon nanotubes, graphene, and their derivatives. 91. M. M. Lucchese, F. Stavale, E. H. M. Ferreira, C. Vilani, M. V. O. Moutinho, R. B. Capaz, C. A. Achete and A. Jorio, Carbon, 2010, 48, 1592-1597. Quantifying defects in graphene via Raman spectroscopy at different excitation energies. 92. M. Zhang, M. Yang, C. Bussy, S. Iijima, K. Kostarelos and M. Yudasaka, Nanoscale, 2015, 7, 2834-2840. Biodegradation of carbon nanohorns in macrophage cells. 93. V. E. Kagan, N. V. Konduru, W. Feng, B. L. Allen, J. Conroy, Y. Volkov, I. I. Vlasova, N. A. Belikova, N. Yanamala, A. Kapralov, Y. Y. Tyurina, J. Shi, E. R. Kisin, A. R. Murray, J. Franks, D. Stolz, P. Gou, J. Klein-Seetharaman, B. Fadeel, A. Star and A. A. Shvedova, Nat. Nanotechnol., 2010, 5, 354-359. Carbon nanotubes degraded by neutrophil myeloperoxidase induce less pulmonary inflammation. 94. N. Lu, J. Li, R. Tian and Y.-Y. Peng, Chem. Res. Toxicol., 2014, 27, 1070-1077. Binding of human serum albumin to single-walled carbon nanotubes activated

271

neutrophils to increase production of hypochlorous acid, the oxidant capable of degrading nanotubes. 95. M. J. Osmond-McLeod, C. A. Poland, F. Murphy, L. Waddington, H. Morris, S. C. Hawkins, S. Clark, R. Aitken, M. J. McCall and K. Donaldson, Part. Fibre Toxicol., 2011, 8, 15. Durability and inflammogenic impact of carbon nanotubes compared with asbestos fibres. 96. T. Burkholder, C. Foltz, E. Karlsson, C. G. Linton and J. M. Smith, Current protocols in mouse biology, 2012, 2, 145-165. Health evaluation of experimental laboratory mice. 97. OECD, Test No. 412: Subacute Inhalation Toxicity: 28-Day Study, OECD Publishing, 2009.

272

Chapter 7

7. Summary and Final Remarks

7.1. Summary.

The oxidised derivative of graphene, known as graphene oxide (GO) has been receiving increasing attention particularly for its potential applicability to biomedical science1, as well as other fields such as materials science. In recent years, GO has been proposed for applications such as drug delivery, tissue engineering and biomedical imaging, this is as well as anticorrosion coatings2 and membrane based desalinisation strategies3. In order to successfully and safely apply the material in these varied applications, the revealing of its biological interactions and fate is essential4.

One aspect of the materials fate that must be understood is its biodegradability4. Previously, the concept that sp2 hybridised carbon nanomaterials and their oxidised derivatives, including GO, could undergo biodegradation, under physiological or environmental conditions was dismayed on account of the high stability and chemical inertness of the materials graphitic backbone5. This was considered a major caveat for the applicability of the material in biomedical applications, where biopersistence is often linked with negative outcomes. This same concept gleaned concern regarding the potential for these materials to accumulate within the environment. However as of 20085,6 evidence to suggest the contrary has emerged with several studies reporting the enzymatic, cellular and in vivo biodegradability of CNMs including fullerenes7, carbon nanotubes5, carbon nanohorns8 as well as graphene9. The field of CNM biodegradation is still in its infancy and until now the idea that CNMs such as GO are biodegradable is still disputed in the scientific arena. In order to gather more evidence to help mature the science that concerns the degradability of CNMs, this thesis was focused on exploring the unknown in vivo biodegradability of highly characterised and defined GO sheets in animal models. Herein, thoroughly characterised GO materials were administered to different organ systems via different routes relevant to biomedical as well as unintended exposures.

This thesis began with a simple but crucial exploration of the changes that oxidised CNMs undergo when exposed to a defined oxidative medium, namely 1% NaClO (Chapter 5). In order to allow the characterisation of the degradability of GO but also to fill a gap in the literature, we compared the GO degradation kinetics to that of two other oxidised 1D CNMs

273 namely oxidised multiwalled carbon nanotubes (MWNTs) and oxidised carbon nanohorns (CNHs) over a week. It was noticed that all three materials underwent degradative changes to different extents and via different mechanisms. The dynamic of degradation was studied with various microscopic and spectroscopic techniques. For GO, the degradation dynamics were such that holes initially formed in the basal planes of the sheets. This could be attributed to the presence of the highly strained and unstable epoxide groups that populate the basal planes of the material. During this time point however, the Raman signature could still be acquired indicating the presence that the graphene backbone was still intact to some extent. By the fifth day, the oxidative degradation had proceeded such that the characteristic Raman signature of GO could no longer be acquired. Moreover at this time point, the sheet like structure of the material could neither be identified, which confirmed the GO sheets structural degradation. As for the other two materials, oxidised MWNT had been degraded with a greatly slower kinetic compared to GO. This was to the extent that the tubular structure and characteristic features of the oxidised MWNTs Raman signature could still be clearly identified by the seventh day of the reaction. Nonetheless, the material did sustain severe defects as evidenced by Raman spectroscopy, transmission electron microscopy (TEM) and atomic force microscopy (AFM). In fact, it was directly observed that a wall exfoliation mechanism was occurring. The oxidised CNHs on the other hand, degraded with intermediate kinetics. The dahlia like structures were progressively broken down, and then the composite CNHs were degraded as evidenced by Raman spectroscopy, TEM and AFM.

Several conclusions were derived from this study that helped to direct and inform further work. Firstly, GO is a highly degradable CNM that is more prone to degradative changes than the other investigated CNMs under the prevailing oxidative conditions. Secondly during the structural degradation of GO sheets, characteristic spectroscopic changes were observed that were in line with the amorphous trajectory of carbon, described previously10. The spectroscopic data was also supported by microscopic observations which suggested that hole formation is a key feature during the early stages of oxidative degradation. Having proved and learned these key pieces of information, we were now more ready to progress with the overall project.

The fate of GO nanosheets was next examined, following their intravenous administration to a mouse model (C57BL/6). In previous publications11,12 the Nanomedicine Lab has shown that; GO sheets are predominantly excreted via the kidney with no toxicological issues following their intravenous administration. The small fraction of the material that remains within the body however, accumulates within the organs of the monophagocytic system, majorly the liver and the spleen, with little or no accumulation within the lungs. Over 24 hours however, the quantity of material progressively decreased within the liver yet

274 increased in paralell within the spleen11. This made the spleen the organ of highest relevance following the intravenous administration of GO nanosheets.

In this section of the thesis, the toxicological consequences of the accumulation of GO nanosheets within the spleen are examined as well as the materials long term in vivo fate within this organ. It is demonstrated that at the intravenous doses administered, GO nanosheets pose no toxicological consequences for the spleen at the structural and functional level, however at the highest dose examined (10 mg/kg) we did detect some minor immunological irregularities. In terms of the sub organ biodistribution, it was show using immunofluorescence and Raman combinations that the material accumulates within the splenic marginal zone, which lies at the intersection of the red and white pulp. Evidence was provided to suggest that the material is primarily sequestered by marginal zone macrophages. These are a highly specialised subpopulation of macrophages13. The fate of this material within these cells was then followed over nine months. The result suggests that the presence of material decreases gradually, with the greatest loss occurring at nine months. In a concordant fashion, there was also a diminishment of the Raman spectral integrity of the detected materials. In a more refined manner in freshly extracted and homogenised tissues, Raman spectral changes of the spleen accumulated GO were tracked over 9 months. It was demonstrated that the material spectra evolved in a way that closely followed the amorphous trajectory of carbon and the predictions of the Tuinstra-Koenin equation10, as was also observed in Chapter 4. It was inferred that the material was undergoing biodegradation.

In order to directly confirm this inference, a protocol was optimised that allowed the direct observation of the materials subcellular localisation in the spleen, as well as to confirm the identity and track the graphene-based materials’ crystallinity over time using TEM with correlative electron diffraction. Strikingly, we saw the GO sheets visual morphology deformed over time and at 30 days post administration, the material had developed holes, correlating with the results of Chapter 4, that were evidently not present in the starting material. By the end of the 9 months the material was harder to detect, but where it was found, the sheet like morphology was no longer apparent. Using electron diffraction, it was confirmed that the materials crystallinity was lost over time and was largely absent at 9 months. Collectively, these evidences demonstrate that following their intravenous administration (a biomedically relevant route of exposure) under the conditions of our experiment, GO nanosheets undergo biodegradation in the spleen.

275

Since this thesis was aimed to interrogate the in vivo biodegradability of GO, the investigation of the biodegradability of the material in the spleen alone was not sufficient to address this question. This is because the spleen is one of many physiological organs. Therefore to continue the investigation, the author decided to explore the biodegradation of the materials in the brain. Unlike the spleen which forms part of the lymphoid system14, the brain is a major component of the central nervous system15. Moreover, since the biodegradation of GO had already been interrogated following a biomedically relevant exposure (Chapter 5), we decided to complete an experiment that was also relevant to unintended exposure contexts. GO sheets of different controlled lateral dimensions, that were now available in the lab, were administered via the intranasal route. This route, as described in Chapter 6, has the possibility to allow the translocation of some substances including various nanomaterials from the nasal cavities to the brain via unique pathways16. Using several methods including SPECT / CT, γ-counting, ICP-MS and Raman spectroscopy, it was shown that GO sheets can translocate from the nose to the brain in a lateral dimension dependent manner, with us-GO translocating to the greatest extent, in terms of quantity and brain coverage. us-GO had predominantly translocated to the olfactory bulb, but the evidence was also suggestive of further translocation as far as the cerebellum, which is at the back of the brain. However, the absolute quantity of us-GO sheets that had actually translocated to the brain as a whole was diminutive, at less than 0.02% of the administered dose. Using immunofluorescence and correlative Raman mapping, it was shown that the majority of the small fraction of material that had been translocated, resided in close association with microglia, the brains resident macrophages. The brain biodistribution of this material was then followed using ICP-MS. It was noticed that the quantity of material in the brain remained constant and so the fate of this material was imperative to track. Raman spectroscopy was therefore employed as previously described in Chapter 5, and the Raman spectra of GO present in freshly extracted brain tissues, of GO exposed mice, were scrutinised to determine if biodegradation of the material in the brain was in part responsible for this decrease. Raman spectra were acquired that evidenced characteristic spectral changes which were suggestive of GOs biodegradation over a month, in a similar manner as previously observed in Chapters 4 and 5. This study was then repeated using a multiple exposure protocol and surprisingly obtained similar results. This ultimately suggests that following the intranasal administration of us-GO to mice models, the brain exposure risk is low. However for the quantity that does translocate to the brain, this undergoes a biodegradative process. Further experiments are needed to understand the products of the biodegradative process and their toxicological significance. Although in the present study no untoward mice weight changes or abnormal behaviours were observed.

276

7.2. Conclusions and further work.

The biodegradability of CNMs is a contentious subject, which is highly debated in science. This is due to the highly stable nature of the graphitic lattice, which is characteristic of these materials5. However as of recent, studies have emerged to suggest the contrary. In this thesis, we have attempted to advance the understanding of the biodegradability of GO, which is among one of the many rising stars of the CNM superfamily particularly with regards to biomedicine1 and material science2,3. The findings of this thesis strongly suggest that the in vivo biodegradation of defined GO materials is possible as we have evidenced with various techniques and in two physiologically distinct organs following both biomedically and unintended relevant exposures. Our results suggest that within these organs, cells of the macrophage lineage are predominantly responsible for this process.

The demonstration of the in vivo biodegradation of GO sheets following biomedically and unintended occupational and private relevant exposures has important implications for the materials long‐term toxicological profile. In terms of biomedicine, this is particularly relevant to any potential clinical application where biodegradability is often an advantage, for example in drug delivery. While with regards to private and occupational health, these findings could help to better understand the health risks posed by these engineered nanomaterials.

Further work should aim to identify the specific molecular mechanisms by which the cells identified here (marginal zone macrophages and microglia, following intravenous and intranasal exposures, respectively) mediate GO biodegradation. This could be completed using techniques such as quantitative polymerase chain reaction and western blotting based techniques as well as mass spectrometry. Moreover, the identity of the biodegradation by- products should be determined; this could be accomplished using mass spectrometry based techniques. The in vivo biodistribution and toxicological implications of these by-products should be examined in the short and long term. Moreover similar degradation studies should be conducted in other organs and following the administration of GO sheets by different routes. For example, further studies should explore the biodegradation of GO in the lungs, which are at high risk following unintended occupational and private exposures particularly via inhalation. Ultimately such further works will help to establish a full understanding of the biodegradability of GO. Furthermore, it will help to inform and lay the foundations for the development of CNMs with tailored degradation kinetics that are suited to different applications.

277

7.3. References.

1 Bitounis, D., Ali-Boucetta, H., Hong, B. H., Min, D. H. & Kostarelos, K. Prospects and challenges of graphene in biomedical applications. Adv. Mater. 25, 2258-2268 (2013). 2 Su, Y. et al. Impermeable barrier films and protective coatings based on reduced graphene oxide. Nat. Commun. 5, 4843 (2014). 3 Abraham, J. et al. Tunable sieving of ions using graphene oxide membranes. Nat Nanotechnol. 12, 546-550 (2017). 4 Kostarelos, K. & Novoselov, K. S. Materials science. Exploring the interface of graphene and biology. Science 344, 261-263 (2014). 5 Kotchey, G. P. et al. A Natural Vanishing Act: The Enzyme-catalyzed degradation of carbon nanomaterials. Acc. Chem. Res. 45, 1770-1781 (2012). 6 Allen, B. L. et al. Biodegradation of single-walled carbon nanotubes through enzymatic catalysis. Nano Lett. 8, 3899-3903 (2008). 7 Litasova, E. V. et al. The biodegradation of fullerene C60 by myeloperoxidase. Dokl. Biochem. Biophys. 471, 417-420 (2016). 8 Zhang, M. et al. Biodegradation of carbon nanohorns in macrophage cells. Nanoscale 7, 2834-2840 (2015). 9 Kotchey, G. P. et al. The enzymatic oxidation of graphene oxide. ACS Nano 5, 2098- 2108 (2011). 10 Ferrari, A. C. & Robertson, J. Interpretation of Raman spectra of disordered and amorphous carbon. Phys. Rev. B 61, 14095-14107 (2000). 11 Jasim, D. A., Menard-Moyon, C., Begin, D., Bianco, A. & Kostarelos, K. Tissue distribution and urinary excretion of intravenously administered chemically functionalized graphene oxide sheets. Chem. Sci. 6, 3952-3964 (2015). 12 Jasim, D. A. et al. The effects of extensive glomerular filtration of thin graphene oxide sheets on kidney physiology. ACS Nano 10, 10753-10767 (2016). 13 Aichele, P. et al. Macrophages of the splenic marginal zone are essential for trapping of blood-borne particulate antigen but dispensable for induction of specific T cell responses. J. Immunol. 171, 1148-1155 (2003). 14 Mebius, R. E. & Kraal, G. Structure and function of the spleen. Nat. Rev. Immunol. 5, 606-616 (2005). 15 Abbott, N. J. & Friedman, A. Overview and introduction: The blood–brain barrier in health and disease. Epilepsia 53, 1-6 (2012).

278

16 Illum, L. Is nose-to-brain transport of drugs in man a reality? J. Pharm. Pharmacol 56, 3-17 (2004).

279