This is the peer reviewed version of the following article: Portioli C., Bussy C., Mazza M., Lozano N., Jasim D.A., Prato M., Bianco A., Bentivoglio M., Kostarelos K.. Intracerebral Injection of Graphene Oxide Nanosheets Mitigates Microglial Activation Without Inducing Acute Neurotoxicity: A Pilot Comparison to Other Nanomaterials. Small, (2020). . : - . 10.1002/smll.202004029, which has been published in final form at https://dx.doi.org/10.1002/smll.202004029. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions. Revised Manuscript 1 1 Injection of Graphene Oxide Nanosheets in the Brain Does 2 3 2 not Induce Acute Neurotoxicity and Counteracts the Acute 4 5 3 Microglial Activation related to Surgery in a Pilot Study 6 7 4 8 9 5 Corinne Portioli1, 2, $, Cyrill Bussy1, 3, 4 *, Mariarosa Mazza1, Neus Lozano1, 3, Dhifaf A. Jasim1,3, 10 11 6 Maurizio Prato5,6, Alberto Bianco7, Marina Bentivoglio2 and Kostas Kostarelos1,3,8, *, & 12 7 13 8 1 Nanomedicine Lab, School of Health Sciences, Faculty of Biology, Medicine & Health, The University of Manchester, 14 9 Manchester Academic Health Science Centre, Manchester, UK 15 16 10 2 Department of Neurosciences, Biomedicine and Movement Sciences, University of Verona, Verona, Italy 17 11 3 National Graphene Institute, The University of Manchester, Manchester, UK 18 12 4 Lydia Becker Institute of Immunology and Inflammation, Faculty of Biology, Medicine and Health, The University of 19 13 Manchester, Manchester Academic Health Science Centre, Manchester, UK 20 21 14 5 Department of Chemical and Pharmaceutical Sciences, University of Trieste, 34127 Trieste, Italy 22 15 6 Carbon Nanobiotechnology Laboratory, CIC BiomaGUNE, 20009 San Sebastian, Spain. 23 16 7 University of Strasbourg, CNRS, Immunology, Immunopathology and Therapeutic Chemistry, UPR 3572, Strasbourg, France 24 17 8 Catalan Institute of Nanoscience and Nanotechnology (ICN2), and The Barcelona Institute of Science and Technology 25 18 (BIST), Campus UAB, Bellaterra, Barcelona, Spain 26 27 19 28 20 29 21 $ Current address: Neuroscience and Brain Technologies Department, Istituto Italiano di Tecnologia (IIT), Genova, Italy and 30 22 Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Baylor College of Medicine (BCM), Houston, TX, 31 32 23 United States 33 24 34 35 25 ORCID numbers: 36 37 26 Cyrill Bussy: 0000-0001-8870-443X 38 39 27 Dhifaf A. Jasim: 0000-0002-6433-4478 40 28 Alberto Bianco: 0000-0002-1090-296X 41 42 29 Marina Bentivoglio: 0000-0003-4958-9795 43 44 30 Kostas Kostarelos: 0000-0002-2224-6672 45 31 46 32 47 48 33 * To whom correspondence should be addressed: 49 34 kostas.kostarelos@manchester.ac.uk ; cyrill.bussy@manchester.ac.uk 50 51 35 & Lead contact: kostas.kostarelos@manchester.ac.uk 52 36 53 54 37 55 56 57 58 59 60 61 62 63 64 65 38 Abstract 1 2 39 Carbon-based nanomaterials (CNMs) are being explored for neurological applications. However, 3 4 40 systematic in vivo studies investigating the effects of CNM nanocarriers in the brain and how brain 5 41 cells respond to such nanomaterials are scarce. To address this, functionalised multi-walled carbon 6 7 42 nanotubes and graphene oxide (GO) sheets were injected in mice brain and compared with charged 8 43 liposomes. The induction of acute neuro-inflammatory and neurotoxic effects locally and in brain 9 10 44 structures distant from the injection site were assessed up to one week post-administration. While 11 45 significant neuronal cell loss and sustained microglial cell activation were observed after injection of 12 13 46 cationic liposomes, none of the tested CNMs induced either neurodegeneration or microglial 14 47 activation. Among the candidate nanocarriers tested, GO sheets appeared to elicit the least 15 16 48 deleterious neuro-inflammatory profile. At molecular level, GO induced moderate activation of pro- 17 49 inflammatory markers compared to vehicle control. At histological level, brain response to GO was 18 19 50 lower than after vehicle control injection, suggesting some capacity for GO to reduce the impact of 20 51 stereotactic injection on brain. While these findings are encouraging and valuable in the selection 21 22 52 and design of nanomaterial-based brain delivery systems, they warrant further investigations to 23 53 better understand the mechanisms underlying GO immunomodulatory properties in brain. 24 25 54 26 27 55 28 56 29 30 57 Keywords 31 32 33 58 Graphene carbon nanotubes liposomes inflammation brain immunomodulation - 34 59 biocompatibility 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 2 62 63 64 65 61 Introduction 1 2 62 Nanomaterials may offer new solutions for unmet medical needs in the treatment of neurological 3 4 63 disorders [1-4]. Among the different types of nanomaterials suitable for these biomedical 5 64 applications, carbon-based nanomaterials (CNMs), including single-walled (SWNTs) or multi-walled 6 7 65 carbon nanotubes (MWNTs) and graphene have recently emerged as potential new candidates given 8 66 their remarkable interaction with the neural tissue [5-10]. CNMs possess unique physicochemical 9 10 67 properties, such as high surface area, mechanical strength, electrical conductivity [11-15] and the 11 68 ability to be chemically functionalised [16, 17]. In the context of neuroscience, these properties have 12 13 69 been shown to support neuronal activity [17] and facilitate drug delivery in the brain [18, 19]. 14 70 15 16 71 Studies performed in vitro have for instance revealed the promising applications of functionalised 17 72 SWNTs as glutamate uptake enhancers in primary astrocytes [20] or as neuroprotective factors in 18 19 73 primary glial cells extracted from brains of an Alzheimer s disease (AD) mouse model [21]. Similarly, 20 74 in animal models, functionalised CNMs were efficient in delivering siRNA in a stroke model [22] or as 21 22 75 drug carrier in an AD model [23]. Another step toward their clinical translation was achieved when 23 76 the translocation of functionalised MWNTs (f-MWNTs) across the blood-brain barrier (BBB) was 24 25 77 demonstrated, initially in vitro [24, 25] and then in vivo [26, 27]. These seminal studies have paved 26 78 the way toward the targeted delivery of active therapeutics across the BBB after peripheral 27 28 79 administration of CNMs, as proposed in one proof-of-concept in vivo study for brain glioma [19]. More 29 80 recently, graphene-based materials (GBMs) and in particular Graphene Oxide (GO), the oxidised 30 31 81 form of graphene that results from chemical exfoliation of graphite, have also been explored for brain 32 82 therapy [6, 15]. Noticeably, GBMs were shown to inhibit the formation of β-amyloid aggregates and 33 34 83 could thus be beneficial in preventing the progression of AD [28]. Then, chemically functionalised GO 35 84 sheets were reported to be suitable photothermal platforms for destroying formed amyloid 36 37 85 aggregates in AD model upon near-infrared light irradiation, via the generation of localised heat [29, 38 39 86 30]. Finally, GBMs were used as nanocarriers for anti-tumour drugs in both in vitro and in vivo 40 87 models of brain cancer [18, 31], and as neurotransmission modulator with potential applications in 41 42 88 neurobiology [32]. 43 89 44 45 90 However, a key issue for a more widespread use of nanocarriers (including CNMs) in brain therapy is 46 91 the response of the brain parenchyma once nanomaterials interact with the different cell populations 47 48 92 of the central nervous system (CNS). This becomes especially crucial in view of potential 49 93 applications of nanocarriers in brain diseases with an inherent neuro-inflammatory component, such 50 51 94 as neurodegeneration, stroke, infection or cancer [33-35]. Therefore, to support the exploration of the 52 95 full potential of CNMs for brain therapy applications, increasing effort has been devoted to investigate 53 54 96 the possible side effects of these materials upon interaction with the brain parenchyma. MWNTs 55 97 coated with polymeric material (Pluronic F127, used to increase solubility of MWNTs) were initially 56 57 98 incubated with primary cortical neurons [36]. As these MWNTs did not induce apoptotic effects in 58 99 vitro, their biocompatibility was then validated in vivo upon injection in the visual cortex of mice [36]. 59 60 100 Similarly, no major tissue damages were reported in another study performed to analyse the neuro- 61 3 62 63 64 65 101 inflammation and cellular uptake of two types of f-MWNTs (carboxylated or amino-functionalised), 1 102 after injection in the cerebral cortex [37]. Both f-MWNT types were internalised by microglial cells and 2 3 103 neurons, and elicited a higher glial cell marker expression at the injection site, 2 days after injection 4 104 [37]. However, at 30 days post-injection, only carboxylated MWNTs resulted in persistent glial cell 5 6 105 activation in regions peripheral to the injection site [37]. In another set of studies, after the infusion of 7 106 PEGylated SWNTs in the hippocampus of rats, an antioxidant response was observed after 24 h [38] 8 9 107 and up to 7 days [39]. The authors theorised that the antioxidant response to SWCNTs could partly 10 11 108 explain the moderate impact of the nanomaterials on animal behaviours [38]; moreover, the 12 109 biopersistence of these CNMs at the injection site was ascribed for the persistence of the antioxidant 13 14 110 response over 7 days [39]. Lastly, a study on the neurotoxic effect of different f-MWNTs using 15 111 primary cultures of neuronal and glial cells derived from either the striatum or frontal cortex revealed 16 17 112 that while f-MWNTs did not affect neuronal cells from any of the two brain regions or glial cells from 18 113 the frontal cortex, the viability of striatum-derived glial cells decreased [40]. Although the brain 19 20 114 region-dependent cytotoxicity to glial cells was shown to be independent of the f-MWNT type, it was 21 115 instead associated with the number of microglial cells in the considered brain region-derived cell 22 23 116 cultures [40], highlighting the key role of microglial cells (the resident macrophages of the brain) in 24 117 the regulation of the biological response to CNMs. 25 26 118 27 119 More recently, the potential impact of GBMs on brain cells and tissue has also been explored. 28 29 120 Functionalised graphene-based systems investigated as drug delivery carriers in the treatment of 30 121 subarachnoid haemorrhage did not show neurotoxicity in the targeted region [41]. However, GO 31 32 122 sheets were reported to down-regulate neuronal activity and signalling in vitro, albeit without affecting 33 123 viability [42, 43]. Autophagy and calcium homeostasis were also found to be disturbed in neuron 34 35 124 cultures exposed to GO, highlighting the ability of GO sheets to damage neuronal transmission and 36 125 functionality, without inducing toxicity [43]. Astrocyte function and homeostasis were similarly altered 37 38 126 by GO sheet exposure and internalisation, subsequently impacting the neuronal network that 39 127 astrocytes were supporting [44]. Finally, when primary mixed glia or the microglia BV2 cell line were 40 41 128 pre-treated with GO sheets, inhibition of NLRP3 inflammasome-dependent interleukin (IL)-1β 42 43 129 secretion was observed upon lipopolysaccharide (LPS) and ATP priming [45]. 44 130 45 46 131 Despite this growing knowledge and the great potential of CNMs as brain drug delivery vectors, 47 132 systematic studies assessing the inflammatory potential of these nanocarriers in brain tissue remain 48 49 133 scarce. To address this gap, three different types of engineered CNMs, including one GO type and 50 134 two f-MWNTs (aminated or carboxylated), were here injected stereotactically into the striatum of mice 51 52 135 and their potent inflammation was assessed. For comparison, two types of highly charged liposomes 53 136 were used as benchmark drug delivery systems with previously reported tissue [46, 47] and brain 54 55 137 [48, 49] inflammogenicity. Considering recent findings highlighting the immunomodulatory and anti- 56 138 inflammatory properties of GO sheets in vitro and in vivo [32, 45, 50], the present study was also 57 58 139 designed to test the hypothesis that GO materials present a unique inflammation profile when 59 140 compared to other nanomaterials. The inflammatory potential of the different candidate nanocarriers 60 61 4 62 63 64 65 141 was therefore assessed at different time points of the acute early stage response (up to 1 week after 1 142 injection) at both the molecular (i.e. transcripts encoding a panel of cytokines and chemokines) and 2 3 143 histological (i.e. activation of astrocytes and microglial cells, number of neurons and dead cells) 4 144 levels. These analyses were performed not only at the injection site (central position in the striatum) 5 6 145 but also in adjacent and distant positions within the brain, to assess both the diffusion of inflammation 7 146 processes and the delocalised effects caused by nanomaterial diffusion. 8 9 147 10 11 148 Results 12 13 149 Characterisation of the NMs 14 150 Either aminated or carboxylated f-MWNTs that have been previously explored for biomedical 15 16 151 applications were used in the present study [22, 37, 51-54]. Their chemical functionalization is 17 18 152 thought to not only improve solubility, but also increase biocompatibility by reducing toxicity through 19 153 mitigation of the material-cell membrane interaction. The dimensional features (diameter and length) 20 21 154 of those f-MWNTs were analysed by transmission electron microscopy (TEM; Figure 1-B, and S1- 22 155 A). Both types of f-MWNTs had an outer diameter between 20 and 30 nm. Carboxylated f-MWCNTs 23 24 156 (ox-MWNTs) were between 200 and 300 nm in length, while aminated f-MWNTs (MWNT-NH3+) had 25 157 a length between 500 nm and 2000 nm. The Kaiser test was performed to establish the amount of 26 27 158 amino groups present on the MWNT-NH3+, and found a loading of 58 µmol/g of amino functional 28 159 groups (Figure S1-A); while the amount of carboxyl group on the ox-MWNTs had been previously 29 30 160 determined using thermogravimetric analysis and found a loading of 1,7 μmol/g [55]. 31 161 32 33 162 In line with our previous works, several techniques were used to assess the physicochemical 34 163 properties of GO sheets (Figures 1-B and S1-B). The ζ-potential was -50.0 ± 0.4 mV. The lateral 35 36 164 dimensions were established with TEM and were in between 10 and 1800 nm, while atomic force 37 165 microscopy (AFM) revealed a thickness between 0.9 and 4.8 nm, consistent with few layer 2D 38 39 166 materials, as we previously reported [56, 57]. 40 167 41 42 168 Characterisation of cationic (DOTAP:Chol) and anionic (DOPG:Chol) liposomes was performed to 43 169 confirm their hydrodynamic diameter size, polydispersity index (PDI) and ζ-potential (Figures 1-B 44 45 170 and S1-C). Cationic (DOTAP:Chol) and anionic (DOPG:Chol) liposomes showed a hydrodynamic 46 171 diameter of 125.6 ± 2.6 nm and 118.1 ± 3.0 nm, respectively. DOTAP:Chol liposomes showed a PDI 47 48 172 of 0.254 ± 0.004, while in the case of DOPG:Chol liposomes, the PDI was 0.393 ± 0.061. The 49 50 173 surface charge of the liposomes was confirmed by ζ-potential measurements. DOTAP:Chol 51 174 liposomes were formed by positively charged polar chains (DOTAP, Figure S1-C) that attribute the 52 53 175 cationic nature to the system (ζ= +60.5 ± 2.6 mV), while DOPG:Chol liposomes were formed by 54 176 negatively charged polar chains (DOPG, Figure S1-C) that attribute the anionic nature to the system 55 56 177 (ζ= -54.1 ± 0.5 mV). 57 178 58 59 179 60 61 5 62 63 64 65 180 Expression of inflammation-related genes 1 2 181 The gene expression levels of transcripts encoding inflammatory molecules were measured in the 3 182 sampled brain tissue blocks. Transcripts encoding TNF-α, IL-1β, IFN-γ, IL-6 and IL-12 were used to 4 5 183 evaluate pro-inflammatory cytokines, CCL2 and CXCL10 as pro-inflammatory chemokines, and IL- 6 184 10, IL-4 and TGF-β as anti-inflammatory markers (Table S1). 7 8 185 9 186 Central brain injection site 10 11 187 The gene expression results for inflammatory markers in the injection site (central striatum) are 12 13 188 presented in Figure 2. As expected, bacterial LPS injection (positive inflammation control) induced 14 189 significantly higher expression levels for all inflammatory transcripts tested, except for ifn-γ mRNA at 15 16 190 day 1 and cxcl10 mRNA at day 1 and 2. At day 7, the upregulation of inflammatory transcripts 17 191 induced by LPS was lower than at the two shorter time points, but remained significantly different 18 19 192 from il-10 mRNA induced by dextrose injection. Surprisingly, there was no significant upregulation of 20 193 cxcl10 expression at any time point. 21 22 194 23 195 In contrast, carbon nanomaterials had a limited effect on the expression levels of these genes 24 25 196 (Figure 2). Over time, ox-MWNTs had a limited impact at day 1 (upregulation of tnf-α and il-1β 26 197 mRNAs), high impact at day 2 (upregulation of il-12, ifn-γ, il-6 and tgf-β mRNAs) and returned to 27 28 198 basal levels at day 7. Similarly, GO upregulated only tgf-β expression at day 1, upregulated tnf-α and 29 199 il-6 expression at day 2, but had no effect at day 7. MWNT-NH3+ upregulated only the il-6 gene at 30 31 200 both day 1 and 2 but had no effect at day 7. Comparison of the three carbon NMs revealed that 32 201 MWNT-NH3+ had the safest inflammatory profile at day 1 while GO was the safest at day 7. At day 2, 33 34 202 both MWNT-NH3+ and GO behaved similarly, while ox-MWNTs induced the greatest inflammation. 35 203 36 37 204 Liposomes were used here as positive nanomaterial controls and were found to more broadly affect 38 205 gene expression (Figure 2). DOTAP:Chol upregulated il-6 and il-10 mRNAs at day 1; this 39 40 206 upregulation persisted at day 2, when expression levels of tnf-α, il-1β and tgf-β mRNAs were also 41 207 upregulated. In addition, the inflammation induced by DOTAP:Chol was maintained at day 7 with 42 43 208 upregulation of ifn-γ, ccl2 and cxcl10 gene transcripts. Similarly, DOPG:Chol upregulated il-12, il-6 44 209 and il-10 mRNAs at day 1 and upregulated tnf-α, il-1β, il-12, il-10 and ccl2 expression at day 2 but 45 46 210 returned to basal level at day 7. When comparing the two types of liposomes, no significant 47 48 211 differences were observed for any inflammatory marker at day 1. At day 2, significant differences 49 212 were found for ccl2 mRNA only. At day 7, significant differences were found for tnf-α, ifn-γ, il-12, ccl2 50 51 213 and cxcl10 mRNAs, revealing an accentuated pro-inflammatory profile for cationic DOTAP:Chol 52 214 liposomes in comparison to anionic DOPG:Chol liposomes. 53 54 215 55 216 Among the different NMs, carbon NMs appeared to elicit the mildest inflammatory response at the 56 57 217 injection site. Both MWNT-NH3+ and GO yielded similar results, whereas ox-MWNT was the most 58 218 pro-inflammatory NM, especially at day 1 and 2 post-injection. 59 60 219 61 6 62 63 64 65 220 Adjacent posterior brain region 1 2 221 The results of gene expression for different inflammatory markers in the posterior brain region in 3 222 direct contact with the injection site are presented in Figure S2. After LPS injection, upregulation of 4 5 223 transcript levels for all markers followed the same trends as in the site of injection. At day 1 and 2, all 6 224 transcripts were upregulated except for ifn-γ mRNA at day 1. At day 7, expression of the tnf-α and il- 7 8 225 1β genes were still upregulated. 9 226 10 11 227 At day 1, carbon NMs had no significant impact on the expression of any of the genes tested in this 12 228 brain region (Figure S2). At day 2, all three carbon NMs regardless of their characteristics 13 14 229 significantly downregulated tnf-α expression and upregulated ccl2 expression. At day 7, none of the 15 230 carbon NMs had any significant effect. No significant differences were observed at any time point 16 17 231 among the three carbon NMs, despite a trend suggesting a mild (compared to nanotubes) 18 232 inflammatory profile after GO administration, especially at day 7 (i.e. ccl2, il-12, and ifn-γ mRNAs had 19 20 233 lower values, albeit without statistical significance). 21 234 22 23 235 In contrast, following injection of liposomes, DOPG:Chol significantly upregulated ccl2 and cxcl10 24 236 mRNAs at day 1, whereas DOTAP:Chol had no effect in the posterior brain region (Figure S2). At 25 26 237 day 2, while DOTAP:Chol significantly upregulated il-1β and il-10 transcripts, DOPG:Chol 27 238 upregulated tnf-α and ccl2 mRNAs. At day 7, none of the liposomes had any effect on inflammatory 28 29 239 marker gene expression, highlighting the transient inflammatory impact of these materials, possibly 30 240 due to their well-known poor long-term structural stability in living tissue. 31 32 241 33 34 242 Distant anterior brain region 35 243 Gene expression levels for the inflammatory markers in the anterior brain region (distant from the 36 37 244 injection site) are presented in Figure S3. As described above, at day 7 in the posterior brain region 38 245 (in direct contact with the injection site), a drastic decrease of the inflammatory response for all 39 40 246 markers and conditions tested was observed, including LPS injection (Figure S2). We therefore 41 247 reasoned that in a distant brain site (not in direct contact with the site of injection) inflammation levels 42 43 248 at day 7 would be even lower. This led us to investigate gene transcripts in the distant anterior brain 44 249 region only at day 1 and 2 (Figure S3). A second motivation for performing analyses of gene 45 46 250 transcripts in the anterior striatum after injection in the middle/central striatum (these two parts of the 47 251 striatum being at relative distance from each other) was brought about the hypothesis that liposomes 48 49 252 can diffuse across this brain region and therefore induce inflammation beyond the site of injection 50 253 [58, 59]. In addition, analyses were performed only for liposomes, as they were inducing upregulation 51 52 254 of genes in the posterior brain region (Figure S2), whereas all carbon NMs did not induce any gene 53 255 upregulation in this brain region (Figure S2). 54 55 256 56 57 257 The results following LPS injection in the anterior brain region were identical to those found for the 58 258 posterior brain region at day 1 and 2, with an upregulation of all markers except for ifn-γ expression 59 60 259 at day 1 (Figure S3). At day 1, anionic DOPG:Chol liposomes elicited the greatest inflammatory 61 7 62 63 64 65 260 response, with significant upregulation of tnf-α, il-1β, il-6, ccl2 and cxcl10 mRNAs, whereas cationic 1 261 DOTAP:Chol only upregulated il-1β expression. In contrast, DOTAP:Chol liposomes were more 2 3 262 inflammatory at day 2, with upregulation of tnf-α, il-1β, ccl2 and cxcl10 expression, while DOPG:Chol 4 263 liposomes only upregulated il-12 expression. 5 6 264 7 265 Overall, anionic DOPG:Chol liposomes seemed to have a greater inflammatory potential at day 1 and 8 9 266 2 not only at the injection site but also in nearby and distant regions of the brain. In contrast, cationic 10 11 267 DOTAP:Chol liposomes showed a greater inflammatory potential at day 2 in all brain regions, 12 268 persisting at day 7 only in the site of injection. These results suggested that liposomes, as 13 14 269 hypothesised, can diffuse across the brain tissue from the injection site and mediate pro- 15 270 inflammatory effects along their path. 16 17 271 18 272 Impact on microglial cells and astrocytes 19 20 273 To investigate the effect of the tested NMs on microglial cells and astrocytes, we focused our efforts 21 274 on day 2. This time point was selected based on the molecular findings presented above, which 22 23 275 indicates that expression levels of pro-inflammatory transcripts were higher 2 days after injection 24 25 276 than at the other time points. The same three brain regions assessed for the RT-qPCR analyses 26 277 were used for the histology study (Figure 1-A). 27 28 278 29 279 Glial cell analyses were based on CD11b and GFAP immunophenotyping. Both qualitative 30 31 280 observations of cell features to detect structural changes indicating an activated state (such as cell 32 281 body hypertrophy and increased thickness of processes) and quantitative analyses were performed 33 34 282 (CD11b, Figure 3 -Ai, -Bi, and -Ci; GFAP, Figure 3 -Aii, -Bii, and -Cii). The latter evaluated the 35 283 following different parameters: i) area covered by microglia and astrocytes, including cell branches (a 36 37 284 higher area indicating cell hypertrophy) (Figures S4-A and S5-A), ii) intensity of microglial cell 38 285 immunoreactivity evaluated by densitometry (increased intensity indicating CD11b upregulation) 39 40 286 (Figure S4-B) and iii) astrocyte cell number (an increased cell number indicates astrocytic activation) 41 287 (Figure S5-B). 42 43 288 The area covered by microglial cells (Figure 3 -Ai, -Bi, and -Ci) and astrocytes (Figure 3 -Aii, -Bii, 44 289 and -Cii) was evaluated for all conditions tested and in the three brain regions considered. The 45 46 290 analysis was also conducted in matched regions of the contralateral hemisphere to obtain data in 47 291 tissue devoid of mechanical trauma due to stereotactic injection or surgery. 48 49 292 50 51 293 Microglial cells 52 294 Immunolabelled microglial cells at the injection site did not show features of activation, or only mild 53 54 295 activation in comparison to vehicle control (5% dextrose in water), at day 2 after injection (Figures 55 296 S4-A and 3-Bi). This was observed for all NMs and doses tested here. Only injection of cationic 56 57 297 DOTAP:Chol or anionic DOPG:Chol liposomes replicated the features observed after LPS injection 58 298 (i.e. hypertrophy and bushy appearance of microglial cells). However, this was not significantly 59 60 299 different from the vehicle control, as analysed quantitatively (Figure 3-Bi). High cell death, likely 61 8 62 63 64 65 300 involving both neurons and glia, was observed at the site of injection after either of these liposomal 1 301 treatments and could account for this finding. Surprisingly, GO injections at 1 mg/ml resulted in 2 3 302 significantly lower CD11b immunoreactivity than control vehicle injection and was similar to that 4 303 observed in the contralateral non-injected hemisphere (Figure 3-Bi). Administration of either type of 5 6 304 f-MWNTs at both 0.5 and 1 mg/ml and GO at 0.5 mg/ml induced a glial cell activation comparable to 7 305 that observed after vehicle injection. 8 9 306 10 11 307 In brain tissue sections distant from the injection site (i.e. anterior brain region), weak microglial cell 12 308 activation was observed after NM injections, while LPS induced clear microglial cell activation 13 14 309 (Figure S4-A). When analysed quantitatively and in comparison to the contralateral region, only LPS 15 310 injection elicited a significant activation of CD11b-positive cells (Figure 3-Ai). All the other conditions, 16 17 311 including both types of f-MWNTs or liposomes and GO, induced microglial cell activation at a level 18 312 similar to that induced by vehicle injection or even lower, and was similar to that observed in the 19 20 313 contralateral hemisphere. Only cationic DOTAP:Chol liposomes (and to a lesser extent DOPG:Chol 21 314 liposomes) induced microglial cell activation that was slightly more pronounced (but not significant) 22 23 315 than the vehicle. 24 316 25 26 317 In sections from brain tissue in direct contact with the injection site (i.e. posterior brain region), mild 27 318 activation of microglial cells was observed after injection of LPS and anionic DOPG:Chol liposomes; 28 29 319 this was also observed to a lesser extent after cationic DOTAP:Chol liposome injection (Figure S4- 30 320 A). Accordingly, quantitative evaluation of the percentage of the area covered by CD11b-positive 31 32 321 cells showed a significant increase only in brains after LPS or DOPG:Chol liposome injection (Figure 33 322 3-Ci). DOTAP:Chol injection induced a modest, but not significant, increase of CD11b-positive cell 34 35 323 coverage compared with both vehicle control injection and the contralateral region. Both types of f- 36 324 MWNTs at either concentration had no effect on activation of microglial cells in this region with 37 38 325 results similar to vehicle control injection. Noticeably, GO injection at either concentration resulted in 39 326 a lower signal than the vehicle and was more comparable to the contralateral region, albeit not 40 41 327 significantly. 42 43 328 44 329 Based on these results, particularly the surprising results obtained with GO versus liposomes or f- 45 46 330 MWNTs, a densitometric evaluation of CD11b immunostaining intensity was performed in brain 47 331 sections containing the injection site and compared against results obtained from LPS- and vehicle- 48 49 332 injected brain tissues (Figure S4-B). Consistent with the findings mentioned above, densitometric 50 333 analysis revealed a significantly lower CD11b optical density after administration of GO than after 51 52 334 vehicle, at the two tested GO doses. These results suggest that the presence of GO could be 53 335 beneficial in reducing the trauma of surgical injection in the striatum. In contrast, and as expected, 54 55 336 CD11b optical density after LPS injection was significantly higher than vehicle injection. Differences 56 337 in this parameter between injection of DOTAP:Chol or DOPG:Chol liposomes and vehicle were not 57 58 338 significant. 59 339 60 61 9 62 63 64 65 340 Astrocytes 1 2 341 Hypertrophic astrocytes, as indicated by higher GFAP immunoreactivity, were observed after LPS 3 342 injection at the injection site (Figure S5-A). Accordingly, the relative surface area covered by 4 5 343 astrocytes after LPS injection showed a significant increase when compared with vehicle control 6 344 injection (Figure 3-Bii). Although not significant, the area covered by GFAP-positive cells was also 7 8 345 slightly increased after injection of ox-MWNT at both concentrations or injection of anionic 9 346 DOPG:Chol liposomes when compared with vehicle. None of the other NMs induced significant 10 11 347 differences when compared with vehicle control injections, but all conditions induced higher astrocyte 12 348 coverage than in the matched, contralateral non-injected brain regions; this suggests that the 13 14 349 mechanical trauma due to their respective injections could account for this mild astrocytic reaction. 15 350 16 17 351 In brain sections from the anterior region (distant from site of injection), mild features of astrocyte 18 352 activation were observed after administration of LPS or cationic DOTAP:Chol liposomes (Figure S5- 19 20 353 A). Accordingly, the relative surface area of the brain tissue covered by GFAP immunoreactivity was 21 354 significantly higher after injection of LPS and DOTAP:Chol liposomes than after vehicle, although 22 23 355 values were lower after injection of DOTAP:Chol than LPS (Figure 3-Aii). None of the other NMs 24 356 induced significant differences. Although not significant, values were lower after GO injection than 25 26 357 after vehicle injection (for both GO doses tested) and were similar to the contralateral hemisphere, 27 358 consistent with the results for microglial cell reactivity after GO treatment. For all conditions, the 28 29 359 relative surface area of the tissue covered by GFAP immunoreactivity was overall lower in this 30 360 anterior brain region than at the injection site (Figure 3-Aii and -Bii). 31 32 361 33 34 362 In the posterior brain region (i.e. sections in close vicinity to the injection site), astrocyte activation 35 363 was observed only after LPS injection and to a far lesser extent after DOTAP:Chol or DOPG:Chol 36 37 364 liposome injection (Figure S5-A). Astrocytes had normal appearance for every other condition. 38 365 These observations were supported by quantitative evaluation of the area covered by GFAP 39 40 366 immunoreactivity (Figure 3-Cii). Only LPS induced a significant increase of GFAP coverage. Values 41 367 after DOTAP:Chol or DOPG:Chol liposome injection were slightly higher than after vehicle 42 43 368 administration, while every other condition showed values similar to or lower than vehicle-injected 44 369 controls. Noticeably, values after GO at 1 mg/ml were lower than after vehicle injection and similar to 45 46 370 those in the matched contralateral brain region. 47 371 48 49 372 Considering these results and the higher astrocyte activation observed in the anterior brain region 50 373 after injection of cationic DOTAP:Chol liposomes, astrocyte cell number was analysed after injection 51 52 374 of DOTAP:Chol liposomes and compared to both positive (LPS) and negative (vehicle) controls 53 375 (Figure S5-B). DOTAP:Chol liposome injection did not significantly affect the number of astrocytes in 54 55 376 the three analysed brain regions, despite being higher than the vehicle control in the anterior brain 56 377 region. The latter result was concordant with the relative coverage of GFAP-positive cells in 57 58 378 DOTAP:Chol liposome-injected brains (Figure 3), which showed greater astrocyte activation than 59 379 after vehicle injection but lower than after LPS injection. Astrocyte number was also significantly 60 61 10 62 63 64 65 380 increased in the anterior and posterior brain regions after LPS injection, but was not significantly 1 381 increased in the injection site. These findings were also in agreement with the relative coverage of 2 3 382 GFAP-positive cells that showed that LPS resulted in higher values in both the anterior and posterior 4 383 brain regions (Figure 3-Aii and -Cii) than in the injection site (Figure 3-E) when compared with 5 6 384 vehicle. As mentioned above, this could be explained by the high cell death (involving glia) elicited by 7 385 LPS in the injection site (Figure 3-Bii). 8 9 386 10 11 387 Impact on neuronal cell viability 12 13 388 Akin to glial cells, the impact of the different NMs on neurons was also studied at day 2 after 14 389 injection. Neuronal cell death following injection of NMs was quantified using NeuN immunostaining 15 16 390 (Figure S6). Statistical evaluation was performed by comparing each data set with that obtained after 17 391 vehicle injection in the respective brain region (Figure 3 -Aiii, -Biii, and -Ciii). Only in the injection 18 19 392 site and after LPS or cationic DOTAP:Chol liposome injections was a significant loss of neurons 20 393 observed (Figure 3-Biii). LPS induced higher neuronal cell loss compared to DOTAP:Chol. No 21 22 394 significant neuronal cell loss was observed after injection of the other NMs or in the other two brain 23 395 regions. 24 25 396 26 397 Based on neuronal cell loss, the number of apoptotic cells were analysed measuring the cleaved 27 28 398 caspase 3 immunoreactivity after LPS and cationic DOTAP:Chol liposome injections, and were 29 399 compared to vehicle and GO injections (Figure 4). The greatest number of cleaved caspase 3 30 31 400 positive cells was found in brain sections containing the injection site after LPS or cationic liposome 32 401 injections, in agreement with the loss of NeuN immunoreactivity. Interestingly, a greater number of 33 34 402 cleaved caspase 3 positive cells was also observed in the cerebral cortex at the level of the injection 35 403 site, possibly associated with the needle track passing through the cortex to reach the striatum, but 36 37 404 only after LPS and DOTAP:Chol liposome injections and not after vehicle or GO injections. This 38 405 suggests a safer toxicological profile for GO than DOTAP:Chol liposomes, consistent with the results 39 40 406 obtained with CD11b immunoreactivity (Figure S3-B). 41 407 42 43 408 Finally, a Fluoro-Jade B staining was performed (Figure S7) to label neurons undergoing 44 409 degeneration [60]. Combining Fluoro-Jade B staining with cleaved caspase 3 staining would help 45 46 410 confirming whether cleaved caspase 3 positive cells were in fact neurons. Consistent with both the 47 48 411 NeuN and cleaved caspase 3 signals, only LPS-injected brains had Fluoro-Jade B labelled neurons 49 412 in the striatum. In the cortical tissue surrounding the needle track, a limited number of Fluoro-Jade B 50 51 413 positive cells were observed after injection of LPS, DOTAP:Chol liposomes, or (to a lesser extent) 52 414 vehicle. Surprisingly, no Fluoro-Jade B stained cells were observed after GO injection in the striatal 53 54 415 or cortical regions, suggesting that the presence of GO sheets may have prevented the brain tissue 55 416 damages associated to surgery and observed after vehicle injection, which supports the data 56 57 417 obtained with CD11b immunoreactivity (Figure S4-B). Results from NeuN and cleaved caspase 3 58 418 immunostaining clearly revealed that NMs inducing neurotoxicity (i.e. DOTAP:Chol liposomes) were 59 60 419 only detrimental at the site of injection, but not in adjacent brain regions. This finding was suggesting 61 11 62 63 64 65 420 that, despite the potency of these materials to diffuse across the brain tissue (as indicated by their 1 421 inflammatory potential across the three tested regions), the amount of diffusing materials is likely to 2 3 422 be limited, or that neurotoxicity requires a high dose of materials, such as found at the injection site, 4 423 to occur. 5 6 424 7 425 Taken together, the results obtained with NeuN, cleaved caspase 3 and Fluoro-Jade B indicate that 8 9 426 only cationic DOTAP:Chol liposomes (and the positive control for inflammation, LPS) had a clear 10 11 427 negative impact on cells and primarily at the site of injection. In contrast, GO appears to have a safer 12 428 and potentially beneficial profile in respect to neurons. 13 14 429 15 16 430 Discussion 17 18 431 Due to their unique properties and dimensions, engineered nanomaterials have emerged as novel 19 20 432 nanomedicine solutions for the treatment or diagnosis of various neurological conditions [1]. 21 433 However, the CNS is a very sensitive environment. If freely bioavailable in the brain parenchyma, 22 23 434 nano-sized foreign materials such as nanocarriers may easily cause disruption to physiological 24 435 processes and functions. It is therefore of greatest importance that safety considerations are 25 26 436 implemented at an early stage during the development of biomedical nanomaterials for CNS 27 437 applications [61]. For this to happen, a better understanding of the nanomaterial physicochemical 28 29 438 characteristics that may induce adverse effects in the brain, such as inflammation, is warranted. This 30 439 is particularly essential for biomedical nanomaterials developed to treat brain diseases that already 31 32 440 have an inflammatory component [34, 35]. 33 441 34 35 442 Recently, both carbon nanotubes and graphene-based materials have shown great promise for the 36 443 treatment and imaging of neurological disorders. However, there is a limited number of studies that 37 38 444 have specifically explored the neuro-inflammation profiles of these CNMs in the brain. With this in 39 445 mind, we went on investigating the neuro-inflammatory potential of different CNMs that could 40 41 446 potentially be used as brain nanomedicines. The tested nanocarriers were directly injected in the 42 43 447 striatum, which was used here as a model of centrally positioned brain region for assessing the 44 448 reactions of the three main cell types of the brain (namely neurons, astrocytes, and microglial cells) 45 46 449 to exogenous materials. Along with CNMs, both cationic and anionic liposomes were used as 47 450 benchmark materials with known inflammatory properties in various tissues [46, 47] or the brain [48, 48 49 451 49, 62]. These inflammatory properties are due to their high density of surface charges. Indeed, while 50 452 anionic micelles were shown to be well tolerated regardless of administration modalities [48], cationic 51 52 453 micelles and cationic liposomes elicited immune cell infiltration and neuronal degeneration due to 53 454 inflammatory response after central administration [48, 49]. 54 55 455 56 456 In the present study, the inflammatory potential of the different nanomaterials was then tested at both 57 58 457 molecular and histological levels. These investigations were performed not only in the area of the brain 59 458 injected with the candidate nanocarriers, but also in adjacent brain areas, either in close vicinity to the 60 61 12 62 63 64 65 459 site of injection (posterior area) or a few mm away from the site of injection (anterior area). This 1 460 assessment in three different locations of the same striatum was designed to assess the possible 2 3 461 diffusion of the materials or biological effects (or both) across the injected brain region, namely the 4 462 striatum. In addition, different doses of nanomaterials were considered. The main tested dose for f- 5 6 463 MWNTs and GO (i.e. 0.5 µg) was based on previous studies for drug delivery purposes using similar 7 464 administration route, bypassing the blood brain barrier [22, 53]. This amount was then doubled to 8 9 465 directly compare with the dose used for liposomes and to assess the role of positive and negative 10 11 466 charges in the inflammation profile of surface-charged CNMs, such as functionalised MWNTs and GO. 12 467 All tested materials were compared to a negative control, an injection with the vehicle (5% dextrose in 13 14 468 water), which reflected the background inflammatory response to the brain tissue damage induced by 15 469 the stereotactic surgical procedure. The reported neuro-inflammation profiles for the different 16 17 470 nanomaterials tested are therefore representing not only the brain tissue response to the material 18 471 injections, but also how each tested nanomaterial modulated the inflammatory response inherent to 19 20 472 the brain surgery used to administer those materials [50]. NM treatments were also compared to LPS, 21 473 a known inflammogenic compound. 22 23 474 24 475 Gene expression analyses of pro- and anti-inflammatory markers revealed that the tested 25 26 476 nanomaterials elicited different patterns of inflammatory response in the considered brain areas. In 27 477 general, regardless of their nature, the levels of pro-inflammatory markers after the administration of 28 29 478 nanomaterials were found to be significantly lower than those elicited by LPS injection at days 1 and 30 479 2, when the LPS-induced upregulation was greatest. But an overall mild acute neuro-inflammatory 31 32 480 response was found for all the different nanomaterials tested, in comparison to the negative control. 33 481 Although the administration of carbon NMs, including GO, elicited a mild upregulation of pro- 34 35 482 inflammatory transcripts immediately after injection, still observable at day 2, gene expression levels 36 483 for these materials were comparable to the negative control by day 7. These findings are in 37 38 484 agreement with previous investigations in which f-MWNTs that were either carboxylated and 39 485 aminated or aminated only, had been injected in the cerebral cortex of mice and induced in both 40 41 486 cases a transient inflammatory reaction, attributed to both nanomaterial and brain surgery, with brain 42 43 487 tissue showing no signs of inflammation by day 14 [37]. Contrastingly, injection of cationic liposomes 44 488 induced marked levels of transcripts encoding pro-inflammatory markers (particularly at the site of 45 46 489 injection) that persisted for up to 7-day after injection, in agreement with previous studies [48, 49]. 47 490 48 49 491 In brain regions close to (but not within) the injection site, there was an overall lower level of pro- 50 492 inflammatory transcripts compared to the injection site. Apart from day 2, the administration of f- 51 52 493 MWNTs and GO did not elicit upregulation of any pro-inflammatory mediators in a nearby brain 53 494 region from the injection site, which is consistent with a previous study that did not reveal any 54 55 495 diffusion of the biological effects after intra-cerebroventricular injection [37]. However, despite the 56 496 distance from the injection site, the administration of cationic liposomes induced marked upregulation 57 58 497 of inflammatory markers in these brain regions at both day 1 and 2. This suggests potential diffusion 59 498 of either the biological response via intercellular signalling, possibly mediated by activated microglial 60 61 13 62 63 64 65 499 cells, or the nanomaterials (or both). This observation is consistent with observations by Knudsen et 1 500 al. [48]. In this study, macrophage infiltration was observed both in the injection site and in a nearby 2 3 501 brain region 1 week after the injection of cationic and anionic liposomes in the dentate gyrus of rats. 4 502 5 6 503 Neurotoxicity leading to neurodegenerative effects represents a major concern for the use of 7 504 nanocarriers in the brain [54]. The results obtained in the present study did not reveal acute neuronal 8 9 505 cell death effects for most of the analysed nanomaterials. Indeed, cell counting revealed a significant 10 11 506 decrease of neuronal cell number in the injection site only after administration of cationic liposomes. 12 507 These findings are consistent with results obtained in a previous study in which the potential 13 14 508 systemic and central toxic responses were evaluated after brain administration of non-PEGylated 15 509 cationic (DOTAP:Chol-Chol) liposomes or PEGylated micelles that were either cationic or anionic 16 17 510 [48]. In the latter study, intra-cerebroventricular administration of cationic liposomes induced 18 511 inflammatory cell infiltration, neuronal degeneration, and cell apoptosis, whereas the administration 19 20 512 of anionic particles did not cause any toxic reaction [48]. Similarly, in the present study, the 21 513 DOPG:Chol anionic liposomes did not induce neurotoxicity, while cationic DOTAP:Chol liposomes 22 23 514 resulted in neurotoxic effects. However, while LPS elicited glial and neuronal cell death at the site of 24 515 injection, cationic liposomes only affected neuron cell number. This could be due to the properties of 25 26 516 cationic liposomes, which are instable nanosystems characterised by rapid clearance due to fusion 27 517 with cell membranes, hence are short-lived [63]. 28 29 518 30 519 In contrast to liposomes, regardless of their surface charge none of the CNMs induced neurotoxicity. 31 32 520 This is consistent with a previous long-term (12 w or 1 y) experimental study on brain tissue 33 521 response following the injection of nanowires with different lengths (2, 5, and 10 μm) in which no 34 35 522 significant differences in the number of neurons were measured [59]. In another study, PEGylated 36 523 SWNTs did not induce cerebral tissue damage or cognitive function alterations at 1 or 7 days after 37 38 524 infusion in the rat hippocampus [39]. In addition, despite short-term oxidative damage observed at 30 39 525 min, an unanticipated antioxidant effect was observed after 7 days, suggesting a potential 40 41 526 neuroprotective ability of these functionalised carbon nanotubes [39]. 42 43 527 44 528 The brain inflammatory response to nanocarrier injection is expected to be mediated by glial cells, 45 46 529 since both microglia and astrocytes act as scavengers for maintaining homeostasis and signalling 47 530 between cells. On one hand, astrocytes control ion and nutrient balance [64], and are activated upon 48 49 531 injury, which manifests structurally by an hypertrophy of the cell body and processes, and an 50 532 upregulation of GFAP [65]. On the other hand, microglial cells are the main CNS immune-resident 51 52 533 components, reacting to early changes in neuronal activity or to pathological conditions [66] and 53 534 constitute the main defence mechanism in the brain. Therefore, the responses of both microglial cells 54 55 535 and astrocytes were analysed in detail in the present study. 56 536 57 58 537 At the injection site, no activation of microglial cells was observed after injection of CNMs, regardless 59 538 of their type or surface charge, whereas hypertrophic microglial cells were observed after injection of 60 61 14 62 63 64 65 539 liposomes (also regardless of their surface charge). Interestingly, in the brain region adjacent to 1 540 injection, mild microglia activation was observed after administration of both positively-charged amino 2 3 541 f-MWNTs and liposomes (anionic or cationic). These findings are consistent with a previous study in 4 542 which a local inflammatory response was induced by f-MWNTs [37]. In addition, a mild astrocyte 5 6 543 response was observed here at the injection site 2 days after injection, particularly after 7 544 administration of ox-MWNTs and anionic liposomes. In adjacent or distant brain regions, the 8 9 545 astrocyte response was significantly lower with respect to the injection site. A previous in vitro study, 10 11 546 performed on primary mixed glial cell cultures, emphasised the importance of microglial cells and 12 547 how their number (with respect to other cells such as astrocytes) can affect biological outcomes [40]. 13 14 548 We observed that administration of cationic liposomes induced astrocyte activation at a distance from 15 549 the injection site, consistent with the pro-inflammatory gene expression response observed in the 16 17 550 same brain region with RT-qPCR analyses. These findings agree well with results obtained in a 18 551 previous study in which both astrogliosis and microgliosis (based on GFAP and Iba1 19 20 552 immunostaining) were identified directly at the site of injection of cationic micelles or in nearby brain 21 553 regions [48]. In contrast, anionic micelles did not induce a similar activation, highlighting the safer 22 23 554 profile of negatively charged nanomaterials when compared to positively charged nanomaterials. 24 555 Indeed, anionic particles interact less with cell membranes that are negatively charged surface. In 25 26 556 contrast, cationic particles, due to a higher electrostatic interaction with negatively charged cells, can 27 557 accumulate to a greater extent in cells and create a more significant burden. This in turn increases 28 29 558 the potential of positively charged nanomaterials to exert a toxic effect [67, 68]. 30 559 31 32 560 Regarding the overall brain inflammation potential of the different nanomaterials tested here, GO 33 34 561 nanosheets appeared to have the least inflammatory profile, when combining both molecular and 35 562 histological results. This is consistent with a recent review that mentioned that, thus far, graphene- 36 37 563 based nanomaterials (including GO) appear to be safer than carbon nanotubes [69]. When 38 564 comparing carbon nanotubes and carbon based two-dimensional lattices, not only the dimensions 39 40 565 (lateral, thickness or length) but also physicochemical features such as rigidity/stiffness or 41 566 bioavailable surface area could be amongst the explanatory material factors making GO more 42 43 567 tolerable than MWNTs under the tested conditions [70, 71]. However systematic investigations 44 568 addressing those questions and comparing the two types of materials are lacking so far, in both the 45 46 569 nanotoxicology and nanomedicine literatures. Here, we observed that GO nanosheets not only 47 570 induced a moderate and acute inflammatory response (tgf-β over-expressed at day 1; tnf-α and il-6 48 49 571 over-expressed at day 2; expression levels similar to negative control for all transcripts by day 7), but 50 572 also led to a lower level of glial cell activation at day 2 when compared to vehicle injection (i.e. glial 51 52 573 activation due solely to surgery in this later case), especially at the 1 mg/mL dose. In addition, GO 53 574 induced less neurotoxicity than LPS, cationic liposomes, or even the vehicle control. This suggests 54 55 575 that the presence of GO in the brain could be beneficial to reduce the impact of intra-parenchymal 56 576 stereotactic surgical injection of materials, a traumatic injury that causes inflammation and cell death 57 58 577 by itself, as evidenced in the negative control results reported here. These findings are consistent 59 578 with a study [50] that reported that mouse brain directly injected with GO had lower GFAP 60 61 15 62 63 64 65 579 immunoreactivity at 48 h and lower iba1 immunoreactivity at 72 h after injection compared to 1 580 negative vehicle control, suggesting that GO had the capacity to lower the activation of astrocytes 2 3 581 and microglial cells, both caused by the brain surgery at the injection site. In agreement with this, 4 582 another study reported the immunomodulatory effects of GO pre-treatment on the macrophage 5 6 583 response to inflammatory challenge [45]. In this study, GO pre-treatment had an anti-inflammatory 7 584 effect upon activation of the inflammasome. Specifically, GO sheets reduced the release of IL-1β and 8 9 585 IL-6 by an NRF2-mediated mechanism. This effect was observed not only in immortalised bone 10 11 586 marrow-derived macrophages but also in a primary murine mixed glia and immortalised microglia 12 587 BV2 cell line. While all these converging findings, including ours, are encouraging from a biomedical 13 14 588 perspective, they warrant further investigations to fully understand the underlying mechanism of the 15 589 immunomodulatory effects of GO nanosheets. In particular a greater sample size, a broader range of 16 17 590 doses and longer time points after injection will be required to reveal how these effects could be 18 591 controlled and safely translated into valuable clinical applications of GO based nanovectors for brain 19 20 592 diseases. 21 22 593 23 594 Conclusion 24 25 595 In the present pilot study looking at the acute response to injection of nanovector candidates in the 26 596 brain, lipid-based NPs, particularly cationic liposomes, induced the greatest inflammatory response in 27 28 597 all considered brain regions. In contrast, CNMs were well-tolerated in the brain parenchyma, with 29 30 598 assessments at both molecular and histological levels revealing only an acute response at days 1 31 599 and 2 followed by fast recovery by day 7. No significant differences were observed between the two 32 33 600 types of MWNT functionalisation or the two doses of CNMs (1 µg vs 0.5 µg). Among the different 34 601 CNMs, GO nanosheets displayed the least deleterious profile, with even some beneficial 35 36 602 immunomodulatory properties that mitigate the inherent inflammation and brain tissue damages 37 603 associated with the brain stereotactic administration. Therefore, under the conditions tested here, GO 38 39 604 nanosheets appeared to have the best profile for future development as brain nanovector, especially 40 605 for cerebral applications that require focal drug administration or in conditions with an inherent 41 42 606 inflammatory component. Going further, additional investigations should examine not only the long- 43 607 term fate and chronic effects of these materials after their injection in the brain, but also the long-term 44 45 608 consequences of the apparent immunomodulation properties of GO. 46 609 47 48 610 Experimental Section 49 50 611 Nanomaterials production 51 52 612 Functionalised multi-walled carbon nanotubes. Pristine MWNTs were purchased from 53 613 Nanostructured and Amorphous Materials Inc. (NanoAmorph, Houston, TX, USA) with a carbon 54 55 614 content of 94%. The pristine materials were then modified using either a 1,3-dipolar cycloaddition 56 615 reaction to obtain aminated MWNTs (MWNT-NH3+) or a 24-h reaction in H2SO4/HNO3 (3:1) solution 57 58 616 to produce carboxylated MWNTs (ox-MWNT), as previously described [72, 73]. 59 617 60 61 16 62 63 64 65 618 Graphene oxide sheets. GO flake suspensions in water were prepared from graphite powder 1 619 (Merck, Sigma-Aldrich, UK) and synthesis was conducted using a modified Hummers method as 2 3 620 previously described [56, 57]. 4 621 5 6 622 Liposomes. To produce liposomes, 1,2-dioleoyl-3-trimethylammonium-propane hydrochloride 7 623 (DOTAP) and 1,2-dioleoyl-sn-glycero-3-[phospho-rac-(3-lysyl(1-glycerol))] (DOPG) were kindly 8 9 624 provided by Lipoid GmbH (Ludwigshafen, Germany). Cholesterol (CHOL) was purchased from Merck 10 11 625 Sigma-Aldrich (UK). Chloroform and methanol were purchased from Thermo Fisher Scientific (UK). 12 626 Both cationic (DOTAP/CHOL; 2 mM DOTAP:1 mM CHOL) and anionic (DOPG/CHOL; 2 mM 13 14 627 DOPG:1 mM CHOL) liposomes were prepared using the film hydration method [74]. Briefly, DOTAP 15 628 or DOPG and CHOL were dissolved in chloroform/methanol (4:1, v/v) and the organic solvents were 16 17 629 evaporated under pressure for 30 min at +40°C using a rotary evaporator. The resulting thin lipid film 18 630 was hydrated in sterile-filtered 5% (w/v) dextrose solution in water and then bath sonicated for 15 min 19 20 631 at +40°C. The final liposome solution was kept at room temperature for 30 min to stabilize the 21 632 colloidal stability before storage at +4°C for a maximum of 5 days. 22 23 633 24 634 Characterization of the nanomaterials 25 26 635 Functionalised multi-walled carbon nanotubes. MWNTs were analysed by transmission electron 27 28 636 microscopy (TEM) to determine the mean diameter and length as previously described for the 29 637 aminated and carboxylated MWNTs [37, 73]. A Kaiser test was used for MWNT-NH3+ to measure the 30 31 638 amount of amine functionalization as previously reported [72]. Kaiser test is based on the colorimetric 32 639 reaction between the ninhydrin reagent and the amine groups. The reaction gives a blue colour 33 34 640 readout and the intensity in proportionally related to the amount of free terminal amine groups [75]. 35 641 DLS was not used to assess hydrodynamic diameter of the GO sheets, as it has been proven non 36 37 642 reliable for one dimensional tube-shaped or two-dimensional plate-shaped materials. More systemic 38 643 characterisation of these materials has been previously reported. 39 40 644 41 645 Graphene oxide sheets. GO sheets were characterised by several techniques, including dynamic 42 43 646 light scattering (DLS, Nano Zeta Sizer ZS, ZEN3600, Malvern Panalytical, Malvern, UK), TEM 44 647 (Philips/FEI, Thermo Fisher Scientific, UK), and atomic force microscopy (AFM, Bruker, UK) to 45 46 648 assess physicochemical properties. These properties include ζ-potential, lateral dimensions, and the 47 649 thickness of the sheets. DLS was not used to assess hydrodynamic diameter of the GO sheets, as it 48 49 650 has been proven non reliable for two-dimensional plate-shaped or one dimensional tube-shaped 50 651 materials. More systemic characterisation of these materials were reported previously [57] (in this 51 52 652 reference, the GO sheets used herein are named small GO). 53 653 54 55 654 Liposomes. Liposomes were first characterised by the DLS technique. Particle diameter and 56 57 655 electrophoretic mobility of cationic and anionic liposomes were measured at 25 ± 0.1°C using a Zeta- 58 656 Sizer unit (Nano Zeta Sizer ZS, ZEN3600, Malvern Panalytical, Malvern, UK). The particle size is 59 60 657 based on DLS in back-scattering mode, at 173° and excitation λ=632.8 nm. For electrophoretic 61 17 62 63 64 65 658 mobility measurements, dispersions were placed into U-shaped cuvettes equipped with gold 1 659 electrodes. The ζ-potential is related to the electrophoretic mobility by Henry s equation valid in the 2 3 660 Smoluchowski approximation, when the screening length is much smaller than the particle radius. 4 661 The prepared liposomes were also analysed using TEM (Philips/FEI, Thermo Fisher Scientific, UK). 5 6 662 7 8 663 Preparation of nanomaterials for brain injection 9 664 Functionalised multi-walled carbon nanotubes. The day before injection, dry powders of MWNT- 10 11 665 NH3+ and ox-MWNT were weighed, exposed to low energy UV light for 6 hours in order to sterilise 12 666 the nanotubes, and then rehydrated with sterile-filtered 5% dextrose solution in water (final 13 14 667 concentration 1 mg/mL) in sterilised glass container. This material suspension was initially sonicated 15 668 for 45 min using a water bath sonicator (VWR, UK) operating at 80 W (45 kHz) to allow dispersion of 16 17 669 the nanotubes in the dextrose solution. A 0.5 mg/mL suspension was achieved by further dilution in 18 670 sterile-filtered 5% dextrose solution. All colloidal suspensions kept at +4°C were sonicated for an 19 20 671 additional 15 min immediately before the injection. 21 672 22 23 673 Graphene oxide sheets. Dry powder of GO sheets that were exposed to UV light for 6 hours after 24 25 674 weighting was also rehydrated in sterile-filtered 5% dextrose solution at a concentration of 1 mg/mL. 26 675 This suspension was sonicated for 30 min using a water bath sonicator (VWR, UK) operating at 80 W 27 28 676 (45 kHz) to allow dispersion of GO flakes in the dextrose solution. The 0.5 mg/mL suspension used 29 677 here was achieved by further dilution in sterile-filtered 5% dextrose solution. 30 31 678 32 679 Liposomes. Liposomes were initially prepared at 2 mM DOTAP:1 mM Chol or 2 mM DOPG:1 mM 33 34 680 Chol and then further diluted to the final concentration of 1 mg/mL in sterile-filtered 5% dextrose 35 681 solution. 36 37 682 38 683 The vehicle used for all nanomaterials, sterile-filtered 5% dextrose solution, was used as negative 39 40 684 control (i.e. basic conditions of inflammation following stereotactic injection of an isotonic solution, 41 685 such as 5% dextrose in water). Lipopolysaccharide (LPS) O111:B4 suspension at 5 mg/mL in sterile- 42 43 686 filtered 5% dextrose solution was used as positive control for inflammatory reaction [37]. 44 687 45 46 688 Animals and sample preparation 47 48 689 A total of 84 young (3-week-old) C57BL/6 male mice were used. The protocol received ethical 49 690 approval from the University of Manchester under authorisation from the United Kingdom Home 50 51 691 Office (project License number PPL-70/7763). Suffering was minimised and the minimal number of 52 692 animals were used in accordance with the Code of Practice for the housing and care of animals used 53 54 693 in scientific procedures. The animals were kept in groups of four to five in standard cages with free 55 694 access to food and water under controlled environmental conditions, including a 12 h/12 h light/dark 56 57 695 cycle. 58 696 59 60 61 18 62 63 64 65 697 For surgery, mice were initially anesthetised with isoflurane inhalation, injected with analgesic 1 698 (Buprenorphine 0.1 mg/kg, im), and then placed on a stereotactic apparatus. A hole was drilled in the 2 3 699 skull at specific lateral coordinates. A total of 1 µl of the different nanomaterials suspended in 5% 4 700 dextrose in water was injected in the striatum with a micro-syringe mounted on a stereotaxic holder 5 6 701 (coordinates used: lateral (x) -0.1 mm, ventral (y) -2.3 mm, rostro-caudal (z) -3.0 mm from bregma) 7 702 [76]. 8 9 703 10 11 704 During the surgical procedure, the mice were kept under oxygen and heated using a blanket with a 12 705 thermostat to maintain body temperature at approximately 37°C. At the end of the procedure, the 13 14 706 wound was sutured and the animal was maintained in a thermally controlled incubation chamber at 15 707 37°C until complete recovery from anaesthesia. The mice were then returned to their maintenance 16 17 708 cages and culled at different time points as shown in the experimental design (Figure 1-A). 18 709 19 20 710 Mice used for gene expression analyses were sacrificed at 1 day, 2 days, or 7 days after injection (n 21 711 = 3 per group; total of 63 mice). They were culled with CO2 exposure followed by cerebral 22 23 712 dislocation. The brain was then rapidly dissected out and cut into 2 mm thick slices using a Zivic 24 713 stainless brain slicer matrix. For each brain, four coronal slices were prepared: one containing the 25 26 714 injected area, one immediately posterior to assess diffusion of nanomaterials or of signal in a region 27 715 adjacent to the injection, and two anterior to the injected area. The most anterior of the latter two 28 29 716 slices was used to assess diffusion of nanomaterials or of signal in a distant brain region. From the 30 717 three brain slices thus sampled, a 2 x 2 x 2 mm tissue block was dissected for RT-qPCR analysis 31 32 718 (Figure 1-A). In the slice containing the injection site, the sampled tissue block was centred on this 33 719 site. In the adjacent posterior slice and in the anterior slice, the tissue blocks were sampled along the 34 35 720 same antero-posterior and dorso-ventral axes of the injection site. The tissue blocks were snap- 36 721 frozen in liquid nitrogen and cryopreserved for RNA extraction and real-time RT-qPCR analysis. 37 38 722 39 723 Animals for colorimetric histochemical and immuno-histochemical procedures were sacrificed at day 40 41 724 2 after injection (n = 3 per group; total of 21 mice). They were anaesthetised by isoflurane inhalation 42 43 725 and then cardiac-perfused with 4% paraformaldehyde in 0.01 M phosphate-buffered saline, pH 7.4 44 726 (PBS). The brain was then dissected out and immersed overnight in 4% paraformaldehyde in PBS. 45 46 727 The following day, brains were soaked in sucrose (5%, 15%, 30% steps) at 4°C for cryoprotection 47 728 following a previously described procedure [37]. 48 49 729 50 730 Real-Time Quantitative PCR analysis 51 52 731 Tissue blocks from animals injected with GO (0.5 mg/mL), MWNT-NH3+ (0.5 mg/mL), ox-MWNT (0.5 53 732 mg/mL), cationic liposomes (1 mg/mL), anionic liposomes (1 mg/mL), 5% dextrose (vehicle, negative 54 55 733 control), or LPS (positive control) were used for transcript analysis. Tissue blocks were homogenised 56 57 734 with a TissueLyser LT (Qiagen, Netherlands) and total RNA was extracted using a NucleoSpin 58 735 RNA/Protein kit (Macherey-Nagel, Germany) according to the manufacturer s instructions. The 59 60 736 concentration of RNA was determined as the optical density ratio 260 nm/280 nm using a 61 19 62 63 64 65 737 BioPhotometer plus (Eppendorf, Germany). Ratio values between 1.8 and 2.2 were considered good 1 738 quality. Samples of cDNA were prepared from 1 mg RNA in a total volume of 20 µl using the BioRad 2 3 739 iScript cDNA Synthesis Kit (BioRad, USA). Samples were run on a CFX-96 Real Time Detection 4 740 System (BioRad, USA) with the following sequence: 95°C for 3 min (initial denaturation step), 1 5 6 741 cycle; 95°C for 10 sec (amplification), 60°C for 30 sec (annealing), repeated for 40 cycles. 7 742 Amplification was followed by a melting-curve analysis to confirm PCR product specificity. 8 9 743 10 11 744 Each RT-PCR reaction in a 25 µl total volume contained 2 µl of cDNA from reverse transcription 12 745 PCR, 12.5 µl Fast SYBR Green Master Mix (BioRad, UK), and primers at 200 nM each (Merck- 13 14 746 Sigma-Aldrich, UK; see Table S1 for reverse and forward primer sequences). Gene expression 15 747 levels (tnf-α, il-1β, il-6, il-12, ifn-γ, cxcl10, ccl2, il-10, tgf-β, il-4, and housekeeping gene β-actin) were 16 17 748 calculated using the Livak method, based on calculation of 2-ΔΔCT [77]. β-actin was used as a 18 749 reference housekeeping gene to normalise the amount of target primer transcripts. The normalised 19 20 750 values for each gene were compared to the relative expression for 5% dextrose (negative control) to 21 751 calculate the fold increase of the target gene in the sample. 22 23 752 24 753 Immuno-histochemical and histochemical procedures 25 26 754 Tissue processing. Brains (day 2 post injection) from animals injected with GO (0.5 and 1 mg/mL), 27 28 755 MWNT-NH3+ (0.5 and 1 mg/mL), ox-MWNT (0.5 and 1 mg/mL), cationic liposomes (1 mg/mL), 29 756 anionic liposomes (1 mg/mL), 5% dextrose (vehicle, negative control), or LPS (positive control) were 30 31 757 used for cell analyses. 32 758 33 34 759 Following cardiac perfusion of fixative under anaesthesia (as described above), post-fixation, and 35 760 brain cryoprotection in sucrose, brains were snap-frozen and then cut using a cryo-microtome into 36 37 761 30-µm-thick serial coronal sections. Series of sections (one every 360 µm) were collected in the 38 762 following three regions: i) anterior to the injection site (from +1.9 to +1.0 from bregma), ii) at the 39 40 763 injection site (from -0.1 to -0.9 from bregma), and iii) posterior to it (from -1.2 to -2.0 from bregma). 41 764 42 43 765 Immuno-phenotyping of neurons, microglia, astrocytes, and apoptotic cells. For each 44 766 experimental group (n = 3 animals per group), a series of sections was processed for immuno- 45 46 767 histochemistry. Free-floating sections were pre-treated with 1% H2O2 (Merck Sigma-Aldrich, UK) for 47 768 15 sec at room temperature, rinsed in PBS (Merck Sigma-Aldrich, UK), and incubated in 5% Normal 48 49 769 Serum of the appropriate species (Vector Lab, USA; Table S2), and 0.03% Triton-X100 (Merck 50 770 Sigma-Aldrich, UK) in PBS for 1 h at room temperature to prevent nonspecific binding. After rinsing in 51 52 771 PBS, the sections were incubated overnight at 4°C in primary antibodies (Table S2) diluted in 1% 53 772 Normal Serum in PBS. The sections were then incubated in biotinylated secondary antibodies 54 55 773 (Vector Lab) in 1% Normal Serum in PBS. The sections were then reacted with the Vectastain ABC 56 57 774 kit (Vector Lab) and finally with 0.5% 3-3 diaminobenzidine (DAB, Merck-Sigma-Aldrich) in PBS. 58 775 After rinsing, the sections were dehydrated through an increasing alcohol gradient, mounted, and 59 60 61 20 62 63 64 65 776 cover-slipped. The sections were examined with an Olympus microscope equipped with a QICAM 1 777 digital camera (QImaging, Canada) using Image-Pro Plus 7.0 Software (Media Cybernetics, USA). 2 3 778 4 779 Fluoro-Jade B histochemistry. To evaluate ongoing neuronal cell death, Fluoro-Jade B staining 5 6 780 was performed [60]. Sections were mounted on gelatin-coated slides, air dried, and soaked for 5 min 7 781 in 1% NaOH (Merck-Sigma-Aldrich) in 80% alcohol in distilled water. The sections were then soaked 8 9 782 for 2 min in 70% alcohol and 2 min in distilled water, and then in a solution of 0.06% potassium 10 11 783 permanganate (Merck Sigma-Aldrich) for 10 min to reduce the background signal. The sections were 12 784 then rinsed in distilled water for 2 min and soaked for 15 min in the staining solution. The Fluoro-Jade 13 14 785 B working solution (0.0004%) was obtained by diluting 4 mL of 0.01% stock solution (10 mg of 15 786 powder [Histochem Inc., USA] in 100 mL of distilled water) into 96 mL of 0.1% acetic acid (Merck 16 17 787 Sigma-Aldrich). The sections were then rinsed in distilled water and air dried. They were cleared in 18 788 xylene for 2 min, mounted, and then cover-slipped. The sections were analysed with an Olympus 19 20 789 microscope equipped with a UV bulb light source (450-490 nm blue excitation light filter; Fluoro-Jade 21 790 B has a green light emission) and images were taken with a QICAM digital camera (QImaging, 22 23 791 Surrey, BC, Canada). 24 792 25 26 793 Quantitative analyses 27 28 794 Counts of neurons and astrocytes. To assess whether the intra-striatal injection of nanomaterials 29 795 induced neuronal cell loss, the number of neurons identified by NeuN immunoreactivity was 30 31 796 estimated using a stereological approach in all groups of mice (animals sacrificed at day 2 after 32 797 injection of all the materials; n = 3 per group; total of 21 mice). 33 34 798 35 799 Stereology was also used to estimate the number of astrocytes. This was based on glial fibrillary 36 37 800 acidic protein (GFAP) immunostaining in the mice treated with 5% dextrose, LPS, or cationic 38 801 liposomes. Cell counting was performed using three regions of interest (ROIs) per section in six 39 40 802 sections (2 regularly spaced sections through the anterior region, the injection site, and the posterior 41 803 region, respectively) per mouse and three mice per group. The counting of astrocytes was performed 42 43 804 in three ROIs per section in six sections sampled as above per mouse and three mice per group. 44 805 Sections were analysed with an Olympus microscope equipped with a Retiga-2000R CCD Camera 45 46 806 (QImaging, Canada) and counting was performed with the Optical Fractionator probe included in 47 807 Stereo Investigator 10 software (MBF Bioscience, USA). 48 49 808 50 809 Analysis of glial cell coverage and optical density of microglial cells. A series of sections for 51 52 810 each condition (n = 3; immunostained as described above for visualisation of microglial cells and 53 811 astrocytes) were used to assess the percentage of the area covered by CD11b- and GFAP-immuno- 54 55 812 positive cells, assuming that a larger area is covered by activated glial cells than by resting cells 56 57 813 [78, 79]. The immunostaining was thus quantified as the percentage of the total image area, 58 814 considering the site of injection in the striatum and an equivalent ROI in the anterior and posterior 59 60 815 sections. 61 21 62 63 64 65 816 1 817 The intensity of the CD11b immunoreactivity was also quantified by densitometry. A quantitative 2 3 818 densitometric analysis [80-82] was performed to measure (in the same sections) the intensity of 4 819 immunoreactivity signal in the cell somata [83]. For this analysis, three ROIs (with an area of 289 5 6 820 μm2) per section in six sections per mouse sampled as above and in three mice per group were 7 821 used. Sections were analysed with an Olympus microscope and 8-bit grey-scale images were taken 8 9 822 with a 20X objective and a QImaging QICAM digital camera (QImaging, Canada) maintaining 10 11 823 constant light conditions and magnification. Images were then processed using the Image-Pro Plus 12 824 7.0 software (Media Cybernetics, USA). A signal from non-immunostained tissue (contralateral 13 14 825 hemisphere) was used to subtract the background signal. 15 826 16 17 827 Statistics 18 828 The results were expressed as mean per group ± standard error of the mean (SEM). The Livak 19 20 829 method was used to analyse qPCR data using ΔCT values [77]. Data were checked for normal 21 830 distribution before running statistical analysis. Statistical variations were evaluated as follows: for 22 23 831 simple comparisons unpaired t-tests were used and one-way analysis of variance (ANOVA) per 24 25 832 group, followed by Bonferroni post-hoc for testing pairwise comparisons. For immuno-histochemical 26 833 and histochemical analysis, the number of sample units used in each study group (n=3) has been 27 28 834 compensated by different measures of the parameter (3 different ROIs) in the areas of interest, within 29 835 the brain (anterior site, injection site, and posterior site). GraphPad Prism (GraphPad Software v.6) 30 31 836 was used for statistical analyses. p-values < 0.05 were considered significant. 32 837 33 34 35 838 Supporting Information 36 839 Supporting Information is available from the Wiley Online Library or from the author. 37 38 840 Pdf document attached. 39 40 841 41 42 842 Funding 43 44 843 This work was partly supported by the EU H2020 RTD Framework Program: FET Graphene Flagship 45 844 project (H2020-EU.1.2.3.- FET Flagships; GrapheneCore; 2785219), the Centre National de la 46 47 845 Recherche Scientifique (CNRS), the International Center for Frontier Research in Chemistry (icFRC), 48 846 and the Agence Nationale de la Recherche (ANR) through the LabEx project Chemistry of Complex 49 50 847 Systems (ANR-10-LABX-0026_CSC). Financial support to this project was partially provided by The 51 848 Fondazione Cariverona ( Verona Nanomedicine Initiative ) and funding from an intramural (University 52 53 849 of Verona) international cooperation program ( CooperInt ) was obtained. KK would like to 54 850 acknowledge the Severo Ochoa Centre of Excellence Award to ICN2. 55 56 851 57 58 852 59 60 853 61 22 62 63 64 65 854 Authors' contributions 1 2 855 CP, CB, and KK conceived the overall design of the project. CP and CB implemented the 3 4 856 experiments and analysed the data under the supervision of CB, KK, and MB, with contributions from 5 857 MM, DAJ, and NL. DAJ prepared the GO sheets and NL prepared the liposomes. AB and MP 6 7 858 provided the carbon nanotubes. CP and CB wrote the manuscript draft. All authors contributed to 8 859 manuscript editing and approved its content. 9 10 860 11 861 Acknowledgments 12 13 862 The authors would like to thank Dr Irene de Lázaro for fruitful discussion regarding statistical 14 15 863 analyses of the RT-qPCR data, Ms Lana Papafilippou for the TEM imaging of the cationic liposomes, 16 864 and Dr Marco Sandri (Big and Open Data Innovation Laboratory - BODaI-Lab, University of Brescia, 17 18 865 Italy) for the valuable scientific support in the biostatistical analysis of the data. Graphical abstract 19 866 and Figure 1A were created using BioRender.com. 20 21 867 22 23 868 Data Availability 24 25 869 The raw/processed data required to reproduce these findings cannot be shared at this time as the 26 870 data also forms part of an ongoing study. 27 28 871 29 30 872 References 31 32 873 [1] D. Furtado, M. Bjornmalm, S. Ayton, A.I. Bush, K. Kempe, F. Caruso, 2018. Overcoming the 33 874 blood-brain barrier: The role of nanomaterials in treating neurological diseases. Adv Mater. 30, 34 875 e1801362. https://doi.org/10.1002/adma.201801362. 35 876 [2] Y. Cheng, R.A. Morshed, B. Auffinger, A.L. Tobias, M.S. Lesniak, Multifunctional nanoparticles for 36 877 brain tumor imaging and therapy, Adv Drug Deliv Rev. 66 (2014) 42-57. 37 878 https://doi.org/10.1016/j.addr.2013.09.006 38 39 879 [3] L. Biddlestone-Thorpe, N. Marchi, K. Guo, C. Ghosh, D. Janigro, K. Valerie, H. Yang, 40 880 Nanomaterial-mediated cns delivery of diagnostic and therapeutic agents, Adv Drug Deliv Rev. 64 41 881 (2012) 605-613. https://doi.org/10.1016/j.addr.2011.11.014. 42 882 [4] A. Dominguez, B. Suarez-Merino, F. Goni-de-Cerio, Nanoparticles and blood-brain barrier: The 43 883 key to central nervous system diseases, J Nanosci Nanotechnol. 14 (2014) 766-779. 44 884 https://doi.org/10.1166/jnn.2014.9119. 45 885 [5] G. Modi, V. Pillay, Y.E. Choonara, V.M. Ndesendo, L.C. du Toit, D. Naidoo, Nanotechnological 46 886 applications for the treatment of neurodegenerative disorders, Prog Neurobiol. 88 (2009) 272-285. 47 887 https://doi.org/10.1016/j.pneurobio.2009.05.002. 48 888 [6] M. Caffo, L. Merlo, D. Marino, G. Caruso, Graphene in neurosurgery: The beginning of a new era, 49 889 Nanomedicine (Lond). 10 (2015) 615-625. https://doi.org/10.2217/nnm.14.195. 50 890 [7] J.T. Wang, K.T. Al-Jamal, Functionalized carbon nanotubes: Revolution in brain delivery, 51 891 Nanomedicine (Lond). 10 (2015) 2639-2642. 10.2217/nnm.15.114. 52 892 [8] H. Kafa, J.T. Wang, K.T. Al-Jamal, Current perspective of carbon nanotubes application in 53 893 neurology, Int Rev Neurobiol. 130 (2016) 229-263. https://doi.org/10.1016/bs.irn.2016.07.001. 54 894 [9] B.S. Wong, S.L. Yoong, A. Jagusiak, T. Panczyk, H.K. Ho, W.H. Ang, G. Pastorin, Carbon 55 895 nanotubes for delivery of small molecule drugs, Adv Drug Deliv Rev. 65 (2013) 1964-2015. 56 896 https://doi.org/10.1016/j.addr.2013.08.005. 57 897 [10] M. Bramini, G. Alberini, E. Colombo, M. Chiacchiaretta, M.L. DiFrancesco, J.F. Maya- 58 898 Vetencourt, L. Maragliano, F. Benfenati, F. Cesca, Interfacing graphene-based materials with neural 59 899 cells, Front Syst Neurosci. 12 (2018) 12. https://doi.org/10.3389/fnsys.2018.00012. 60 61 23 62 63 64 65 900 [11] A. Bianco, K. Kostarelos, C.D. Partidos, M. Prato, Biomedical applications of functionalised 1 901 carbon nanotubes, Chem Commun (Camb). 5 (2005) 571-577. https://doi.org/10.1039/b410943k. 2 902 [12] A. Bianco, K. Kostarelos, M. Prato, Applications of carbon nanotubes in drug delivery, Curr Opin 3 903 Chem Biol. 9 (2005) 674-679. https://doi.org/10.1016/j.cbpa.2005.10.005. 4 904 [13] L. Lacerda, A. Bianco, M. Prato, K. Kostarelos, Carbon nanotubes as nanomedicines: From 5 905 toxicology to pharmacology, Adv Drug Deliv Rev. 58 (2006) 1460-1470. 6 906 https://doi.org/10.1016/j.addr.2006.09.015. 7 907 [14] K. Kostarelos, K.S. Novoselov, Graphene devices for life, Nat Nanotechnol. 9 (2014) 744-745. 8 908 https://doi.org/10.1038/nnano.2014.224. 9 909 [15] T.A. Mattei, How graphene is expected to impact neurotherapeutics in the near future, Expert 10 910 Rev Neurother. 14 (2014) 845-847. https://doi.org/10.1586/14737175.2014.925804. 11 911 [16] X. Guo, N. Mei, Assessment of the toxic potential of graphene family nanomaterials, J Food 12 912 Drug Anal. 22 (2014) 105-115. https://doi.org/10.1016/j.jfda.2014.01.009. 13 913 [17] K. Zhou, G.A. Thouas, C.C. Bernard, D.R. Nisbet, D.I. Finkelstein, D. Li, J.S. Forsythe, Method 14 914 to impart electro- and biofunctionality to neural scaffolds using graphene polyelectrolyte multilayers, 15 915 ACS Appl Mater Interfaces. 4 (2012) 4524-4531. https://doi.org/10.1021/am3007565. 16 916 [18] S.M. Chowdhury, C. Surhland, Z. Sanchez, P. Chaudhary, M.A. Suresh Kumar, S. Lee, L.A. 17 917 Pena, M. Waring, B. Sitharaman, M. Naidu, Graphene nanoribbons as a drug delivery agent for 18 918 lucanthone mediated therapy of glioblastoma multiforme, Nanomedicine. 11 (2015) 109-118. 19 919 https://doi.org/10.1016/j.nano.2014.08.001. 20 920 [19] J. Ren, S. Shen, D. Wang, Z. Xi, L. Guo, Z. Pang, Y. Qian, X. Sun, X. Jiang, The targeted 21 921 delivery of anticancer drugs to brain glioma by pegylated oxidized multi-walled carbon nanotubes 22 922 modified with angiopep-2, Biomaterials. 33 (2012) 3324-3333. 23 923 https://doi.org/10.1016/j.biomaterials.2012.01.025. 24 924 [20] M.K. Gottipati, E. Bekyarova, R.C. Haddon, V. Parpura, Chemically functionalized single-walled 25 925 carbon nanotubes enhance the glutamate uptake characteristics of mouse cortical astrocytes, Amino 26 926 Acids. 47 (2015) 1379-1388. https://doi.org/10.1007/s00726-015-1970-9. 27 927 [21] X. Xue, L.R. Wang, Y. Sato, Y. Jiang, M. Berg, D.S. Yang, R.A. Nixon, X.J. Liang, Single-walled 28 928 carbon nanotubes alleviate autophagic/lysosomal defects in primary glia from a mouse model of 29 929 alzheimer's disease, Nano Lett. 14 (2014) 5110-5117. https://doi.org/10.1021/nl501839q. 30 930 [22] K.T. Al-Jamal, L. Gherardini, G. Bardi, A. Nunes, C. Guo, C. Bussy, M.A. Herrero, A. Bianco, M. 31 931 Prato, K. Kostarelos, T. Pizzorusso, Functional motor recovery from brain ischemic insult by carbon 32 932 nanotube-mediated sirna silencing, Proc Natl Acad Sci USA. 108 (2011) 10952-10957. 33 933 https://doi.org/10.1073/pnas.1100930108. 34 934 [23] Z. Yang, Y. Zhang, Y. Yang, L. Sun, D. Han, H. Li, C. Wang, Pharmacological and toxicological 35 935 target organelles and safe use of single-walled carbon nanotubes as drug carriers in treating 36 936 alzheimer disease, Nanomedicine. 6 (2010) 427-441. https://doi.org/10.1016/j.nano.2009.11.007. 37 937 [24] H. Kafa, J.T. Wang, N. Rubio, K. Venner, G. Anderson, E. Pach, B. Ballesteros, J.E. Preston, 38 938 N.J. Abbott, K.T. Al-Jamal, The interaction of carbon nanotubes with an in vitro blood-brain barrier 39 939 model and mouse brain in vivo, Biomaterials. 53 (2015) 437-452. 40 940 https://doi.org/10.1016/j.biomaterials.2015.02.083. 41 941 [25] S. Shityakov, E. Salvador, G. Pastorin, C. Forster, Blood-brain barrier transport studies, 42 43 942 aggregation, and molecular dynamics simulation of multiwalled carbon nanotube functionalized with 44 943 fluorescein isothiocyanate, Int J Nanomedicine. 10 (2015) 1703-1713. 45 944 https://doi.org/10.2147/IJN.S68429. 46 945 [26] H. Kafa, J.T. Wang, N. Rubio, R. Klippstein, P.M. Costa, H.A. Hassan, J.K. Sosabowski, S.S. 47 946 Bansal, J.E. Preston, N.J. Abbott, K.T. Al-Jamal, Translocation of lrp1 targeted carbon nanotubes of 48 947 different diameters across the blood-brain barrier in vitro and in vivo, J Control Release. 225 (2016) 49 948 217-229. https://doi.org/10.1016/j.jconrel.2016.01.031. 50 949 [27] J.T. Wang, N. Rubio, H. Kafa, E. Venturelli, C. Fabbro, C. Menard-Moyon, T. Da Ros, J.K. 51 950 Sosabowski, A.D. Lawson, M.K. Robinson, M. Prato, A. Bianco, F. Festy, J.E. Preston, K. 52 951 Kostarelos, K.T. Al-Jamal, Kinetics of functionalised carbon nanotube distribution in mouse brain 53 952 after systemic injection: Spatial to ultra-structural analyses, J Control Release. 224 (2016) 22-32. 54 953 https://doi.org/10.1016/j.jconrel.2015.12.039. 55 954 [28] Y. Liu, L.P. Xu, W. Dai, H. Dong, Y. Wen, X. Zhang, Graphene quantum dots for the inhibition of 56 955 beta amyloid aggregation, Nanoscale. 7 (2015) 19060-19065. https://doi.org/10.1039/c5nr06282a. 57 956 [29] M. Li, X. Yang, J. Ren, K. Qu, X. Qu, Using graphene oxide high near-infrared absorbance for 58 957 photothermal treatment of alzheimer's disease, Adv Mater. 24 (2012) 1722-1728. 59 958 10.1002/adma.201104864. 60 61 24 62 63 64 65 959 [30] L. Feng, L. Wu, X. Qu, New horizons for diagnostics and therapeutic applications of graphene 1 960 and graphene oxide, Adv Mater. 25 (2013) 168-186. 10.1002/adma.201203229. 2 961 [31] G. Liu, H. Shen, J. Mao, L. Zhang, Z. Jiang, T. Sun, Q. Lan, Z. Zhang, Transferrin modified 3 962 graphene oxide for glioma-targeted drug delivery: In vitro and in vivo evaluations, ACS Appl Mater 4 963 Interfaces. 5 (2013) 6909-6914. https://doi.org/10.1021/am402128s. 5 964 [32] Y. Kang, J. Liu, S. Yin, Y. Jiang, X. Feng, J. Wu, Y. Zhang, A. Chen, Y. Zhang, L. Shao, 6 965 Oxidation of reduced graphene oxide via cellular redox signaling modulates actin-mediated 7 966 neurotransmission, ACS NANO. (2020). 10.1021/acsnano.9b08078. 8 967 [33] H. Peluffo, U. Unzueta, M.L. Negro-Demontel, Z. Xu, E. Vaquez, N. Ferrer-Miralles, A. 9 968 Villaverde, Bbb-targeting, protein-based nanomedicines for drug and nucleic acid delivery to the cns, 10 969 Biotechnol Adv. 33 (2015) 277-287. https://doi.org/10.1016/j.biotechadv.2015.02.004 11 970 [34] M.W. Salter, B. Stevens, Microglia emerge as central players in brain disease, Nat Med. 23 12 971 (2017) 1018-1027. https://doi.org/10.1038/nm.4397. 13 972 [35] M. Bentivoglio, R. Mariotti, G. Bertini, Neuroinflammation and brain infections: Historical context 14 973 and current perspectives, Brain Res Rev. 66 (2011) 152-173. 15 974 https://doi.org/10.1016/j.brainresrev.2010.09.008. 16 975 [36] G. Bardi, P. Tognini, G. Ciofani, V. Raffa, M. Costa, T. Pizzorusso, Pluronic-coated carbon 17 976 nanotubes do not induce degeneration of cortical neurons in vivo and in vitro, Nanomedicine. 5 18 977 (2009) 96-104. https://doi.org/10.1016/j.nano.2008.06.008. 19 978 [37] G. Bardi, A. Nunes, L. Gherardini, K. Bates, K.T. Al-Jamal, C. Gaillard, M. Prato, A. Bianco, T. 20 979 Pizzorusso, K. Kostarelos, 2013. Functionalized carbon nanotubes in the brain: Cellular 21 980 internalization and neuroinflammatory responses. PLoS One. 8, e80964. 22 981 https://doi.org/10.1371/journal.pone.0080964. 23 982 [38] L. Dal Bosco, G.E. Weber, G.M. Parfitt, K. Paese, C.O. Goncalves, T.M. Serodre, C.A. Furtado, 24 983 A.P. Santos, J.M. Monserrat, D.M. Barros, Pegylated carbon nanotubes impair retrieval of contextual 25 984 fear memory and alter oxidative stress parameters in the rat hippocampus, Biomed Res Int. 2015 26 985 (2015) 1-11. https://doi.org/10.1155/2015/104135. 27 986 [39] L. Dal Bosco, G.E. Weber, G.M. Parfitt, A.P. Cordeiro, S.K. Sahoo, C. Fantini, M.C. Klosterhoff, 28 987 L.A. Romano, C.A. Furtado, A.P. Santos, J.M. Monserrat, D.M. Barros, 2015. Biopersistence of 29 988 pegylated carbon nanotubes promotes a delayed antioxidant response after infusion into the rat 30 989 hippocampus. PLoS One. 10, e0129156. https://doi.org/10.1371/journal.pone.0129156. 31 990 [40] C. Bussy, K.T. Al-Jamal, J. Boczkowski, S. Lanone, M. Prato, A. Bianco, K. Kostarelos, Microglia 32 991 determine brain region-specific neurotoxic responses to chemically functionalized carbon nanotubes, 33 992 ACS NANO. 9 (2015) 7815-7830. https://doi.org/10.1021/acsnano.5b02358. 34 993 [41] L. Yang, F. Wang, H. Han, L. Yang, G. Zhang, Z. Fan, Functionalized graphene oxide as a drug 35 994 carrier for loading pirfenidone in treatment of subarachnoid hemorrhage, Colloids Surf B 36 995 Biointerfaces. 129 (2015) 21-29. https://doi.org/10.1016/j.colsurfb.2015.03.022. 37 996 [42] R. Rauti, N. Lozano, V. Leon, D. Scaini, M. Musto, I. Rago, F.P. Ulloa Severino, A. Fabbro, L. 38 997 Casalis, E. Vazquez, K. Kostarelos, M. Prato, L. Ballerini, Graphene oxide nanosheets reshape 39 998 synaptic function in cultured brain networks, ACS NANO. 10 (2016) 4459-4471. 40 999 https://doi.org/10.1021/acsnano.6b00130. 41 1000 [43] M. Bramini, S. Sacchetti, A. Armirotti, A. Rocchi, E. Vazquez, V. Leon Castellanos, T. Bandiera, 42 43 1001 F. Cesca, F. Benfenati, Graphene oxide nanosheets disrupt lipid composition, ca(2+) homeostasis, 44 1002 and synaptic transmission in primary cortical neurons, ACS NANO. 10 (2016) 7154-7171. 45 1003 https://doi.org/10.1021/acsnano.6b03438. 46 1004 [44] M. Chiacchiaretta, M. Bramini, A. Rocchi, A. Armirotti, E. Giordano, E. Vazquez, T. Bandiera, S. 47 1005 Ferroni, F. Cesca, F. Benfenati, Graphene oxide upregulates the homeostatic functions of primary 48 1006 astrocytes and modulates astrocyte-to-neuron communication, Nano Lett. (2018). 49 1007 https://doi.org/10.1021/acs.nanolett.8b02487. 50 1008 [45] C. Hoyle, J. Rivers-Auty, E. Lemarchand, S. Vranic, E. Wang, M. Buggio, N.J. Rothwell, S.M. 51 1009 Allan, K. Kostarelos, D. Brough, Small, thin graphene oxide is anti-inflammatory activating nuclear 52 1010 factor erythroid 2-related factor 2 via metabolic reprogramming, ACS NANO. 12 (2018) 11949-11962. 53 1011 https://doi.org/10.1021/acsnano.8b03642. 54 1012 [46] H. Lv, S. Zhang, B. Wang, S. Cui, J. Yan, Toxicity of cationic lipids and cationic polymers in gene 55 1013 delivery, J Control Release. 114 (2006) 100-109. https://doi.org/10.1016/j.jconrel.2006.04.014. 56 1014 [47] M.C. Filion, N.C. Phillips, Toxicity and immunomodulatory activity of liposomal vectors 57 1015 formulated with cationic lipids toward immune effector cells, Biochim Biophys Acta. 1329 (1997) 345- 58 1016 356. https://doi.org/10.1016/S0005-2736(97)00126-0. 59 1017 [48] K.B. Knudsen, H. Northeved, P.K. Ek, A. Permin, T.L. Andresen, S. Larsen, K.M. Wegener, H.R. 60 1018 Lam, J. Lykkesfeldt, Differential toxicological response to positively and negatively charged 61 25 62 63 64 65 1019 nanoparticles in the rat brain, Nanotoxicology. 8 (2014) 764-774. 1 1020 https://doi.org/10.3109/17435390.2013.829589. 2 1021 [49] K.B. Knudsen, H. Northeved, P.E. Kumar, A. Permin, T. Gjetting, T.L. Andresen, S. Larsen, K.M. 3 1022 Wegener, J. Lykkesfeldt, K. Jantzen, S. Loft, P. Moller, M. Roursgaard, In vivo toxicity of cationic 4 1023 micelles and liposomes, Nanomedicine. 11 (2015) 467-477. 5 1024 https://doi.org/10.1016/j.nano.2014.08.004. 6 1025 [50] R. Rauti, M. Medelin, L. Newman, S. Vranic, G. Reina, A. Bianco, M. Prato, K. Kostarelos, L. 7 1026 Ballerini, Graphene oxide flakes tune excitatory neurotransmission in vivo by targeting hippocampal 8 1027 synapses, Nano Lett. 19 (2019) 2858-2870. 10.1021/acs.nanolett.8b04903. 9 1028 [51] C. Klumpp, K. Kostarelos, M. Prato, A. Bianco, Functionalized carbon nanotubes as emerging 10 1029 nanovectors for the delivery of therapeutics, Biochim Biophys Acta. 1758 (2006) 404-412. 11 1030 10.1016/j.bbamem.2005.10.008. 12 1031 [52] A. Bianco, K. Kostarelos, M. Prato, Making carbon nanotubes biocompatible and biodegradable, 13 1032 Chem Commun 47 (2011) 10182-10188. 10.1039/c1cc13011k. 14 1033 [53] A. Nunes, C. Bussy, L. Gherardini, M. Meneghetti, M.A. Herrero, A. Bianco, M. Prato, T. 15 1034 Pizzorusso, K.T. Al-Jamal, K. Kostarelos, In vivo degradation of functionalized carbon nanotubes 16 1035 after stereotactic administration in the brain cortex, Nanomedicine (Lond). 7 (2012) 1485-1494. 17 1036 https://doi.org/10.2217/nnm.12.33. 18 1037 [54] A. Nunes, K. Al-Jamal, T. Nakajima, M. Hariz, K. Kostarelos, Application of carbon nanotubes in 19 1038 neurology: Clinical perspectives and toxicological risks, Arch Toxicol. 86 (2012) 1009-1020. 20 1039 https://doi.org/10.1007/s00204-012-0860-0. 21 1040 [55] C. Bussy, C. Hadad, M. Prato, A. Bianco, K. Kostarelos, Intracellular degradation of chemically 22 1041 functionalized carbon nanotubes using a long-term primary microglial culture model, Nanoscale. 8 23 1042 (2016) 590-601. 10.1039/c5nr06625e. 24 1043 [56] A.D. Jasim, N. Lozano, K. Kostarelos, 2016. Synthesis of few-layered, high-purity graphene 25 1044 oxide sheets from different graphite sources for biology. 2D materials. 3, 014006. 26 1045 https://doi.org/10.1088/2053-1583/3/1/014006. 27 1046 [57] A.F. Rodrigues, L. Newman, N. Lozano, S.P. Mukherjee, B. Fadeel, C. Bussy, K. Kostarelos, 28 1047 2018. A blueprint for the synthesis and characterisation of thin graphene oxide with controlled lateral 29 1048 dimensions for biomedicine. 2D materials. 5, 035020. https://doi.org/10.1088/2053-1583/aac05c. 30 1049 [58] C. Mamot, J.B. Nguyen, M. Pourdehnad, P. Hadaczek, R. Saito, J.R. Bringas, D.C. Drummond, 31 1050 K. Hong, D.B. Kirpotin, T. McKnight, M.S. Berger, J.W. Park, K.S. Bankiewicz, Extensive distribution 32 1051 of liposomes in rodent brains and brain tumors following convection-enhanced delivery, J 33 1052 Neurooncol. 68 (2004) 1-9. https://doi.org/10.1023/B:NEON.0000024743.56415.4b. 34 1053 [59] L. Gallentoft, L.M. Pettersson, N. Danielsen, J. Schouenborg, C.N. Prinz, C.E. Linsmeier, Size- 35 1054 dependent long-term tissue response to biostable nanowires in the brain, Biomaterials. 42 (2015) 36 1055 172-183. https://doi.org/10.1016/j.biomaterials.2014.11.051. 37 1056 [60] L.C. Schmued, K.J. Hopkins, Fluoro-jade b: A high affinity fluorescent marker for the localization 38 1057 of neuronal degeneration, Brain Res. 874 (2000) 123-130. https://doi.org/10.1016/S0006- 39 1058 8993(00)02513-0. 40 1059 [61] V. Mirshafiee, W. Jiang, B. Sun, X. Wang, T. Xia, Facilitating translational nanomedicine via 41 1060 predictive safety assessment, Mol Ther. 25 (2017) 1522-1530. 42 43 1061 https://doi.org/10.1016/j.ymthe.2017.03.011. 44 1062 [62] P. Møller, J. Lykkesfeldt, Positive charge, negative effect: The impact of cationic nanoparticles in 45 1063 the brain, Nanomedicine (Lond). 9 (2014) 1441-1443. https://doi.org/10.2217/nnm.14.91. 46 1064 [63] M.L. Immordino, F. Dosio, L. Cattel, Stealth liposomes: Review of the basic science, rationale, 47 1065 and clinical applications, existing and potential, Int J Nanomedicine. 1 (2006) 297-315. 2426795 48 1066 [64] M. Simard, M. Nedergaard, The neurobiology of glia in the context of water and ion homeostasis, 49 1067 Neuroscience. 129 (2004) 877-896. https://doi.org/10.1016/j.neuroscience.2004.09.053. 50 1068 [65] M.V. Sofroniew, H.V. Vinters, Astrocytes: Biology and pathology, Acta Neuropathol. 119 (2010) 51 1069 7-35. https://doi.org/10.1007/s00401-009-0619-8. 52 1070 [66] M.L. Block, L. Zecca, J.S. Hong, Microglia-mediated neurotoxicity: Uncovering the molecular 53 1071 mechanisms, Nat Rev Neurosci. 8 (2007) 57-69. https://doi.org/10.1038/nrn2038. 54 1072 [67] S.J. Soenen, E. Illyes, D. Vercauteren, K. Braeckmans, Z. Majer, S.C. De Smedt, M. De Cuyper, 55 1073 The role of nanoparticle concentration-dependent induction of cellular stress in the internalization of 56 1074 non-toxic cationic magnetoliposomes, Biomaterials. 30 (2009) 6803-6813. 57 1075 https://doi.org/10.1016/j.biomaterials.2009.08.050. 58 1076 [68] R. Banerjee, Liposomes: Applications in medicine, J Biomater Appl. 16 (2001) 3-21. 59 1077 https://doi.org/10.1106/RA7U-1V9C-RV7C-8QXL. 60 61 26 62 63 64 65 1078 [69] B. Fadeel, C. Bussy, S. Merino, E. Vazquez, E. Flahaut, F. Mouchet, L. Evariste, L. Gauthier, 1 1079 A.J. Koivisto, U. Vogel, C. Martin, L.G. Delogu, T. Buerki-Thurnherr, P. Wick, D. Beloin-Saint-Pierre, 2 1080 R. Hischier, M. Pelin, F. Candotto Carniel, M. Tretiach, F. Cesca, F. Benfenati, D. Scaini, L. Ballerini, 3 1081 K. Kostarelos, M. Prato, A. Bianco, Safety assessment of graphene-based materials: Focus on 4 1082 human health and the environment, ACS NANO. 12 (2018) 10582-10620. 5 1083 https://doi.org/10.1021/acsnano.8b04758. 6 1084 [70] C. Bussy, H. Ali-Boucetta, K. Kostarelos, Safety considerations for graphene: Lessons learnt 7 1085 from carbon nanotubes, Acc Chem Res. 46 (2013) 692-701. 10.1021/ar300199e. 8 1086 [71] M. Rezazadeh Azari, Y. Mohammadian, Comparing in vitro cytotoxicity of graphite, short multi- 9 1087 walled carbon nanotubes, and long multi-walled carbon nanotubes, Environ Sci Pollut Res Int. 10 1088 (2020). 10.1007/s11356-020-08036-4. 11 1089 [72] V. Georgakilas, N. Tagmatarchis, D. Pantarotto, A. Bianco, J.P. Briand, M. Prato, Amino acid 12 1090 functionalisation of water soluble carbon nanotubes, Chem Commun (Camb). (2002) 3050-3051. 13 1091 https://doi.org/10.1039/b209843a. 14 1092 [73] S. Li, W. Wu, S. Campidelli, V. Sarnatskaïa, M. Prato, A. Tridon, A. Nikolaev, V. Nikolaev, A. 15 1093 Bianco, E. Snezhkova, Adsorption of carbon nanotubes on active carbon microparticles, Carbon. 46 16 1094 (2008) 1091-1095. https://doi.org/10.1016/j.carbon.2008.03.010. 17 1095 [74] N. Lozano, W.T. Al-Jamal, A. Taruttis, N. Beziere, N.C. Burton, J. Van den Bossche, M. Mazza, 18 1096 E. Herzog, V. Ntziachristos, K. Kostarelos, Liposome-gold nanorod hybrids for high-resolution 19 1097 visualization deep in tissues, J Am Chem Soc. 134 (2012) 13256-13258. 20 1098 https://doi.org/10.1021/ja304499q. 21 1099 [75] E. Kaiser, R.L. Colescott, C.D. Bossinger, P.I. Cook, Color test for detection of free terminal 22 1100 amino groups in the solid-phase synthesis of peptides, Anal Biochem. 34 (1970) 595-598. 23 1101 10.1016/0003-2697(70)90146-6. 24 1102 [76] G. Paxinos, K. Franklin, Paxinos and franklin's the mouse brain in stereotaxic coordinates, ed, 25 1103 2012 26 1104 [77] K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using real-time 27 1105 quantitative pcr and the 2(-delta delta c(t)) method, Methods. 25 (2001) 402-408. 28 1106 https://doi.org/10.1006/meth.2001.1262. 29 1107 [78] H. Dou, K. Birusingh, J. Faraci, S. Gorantla, L.Y. Poluektova, S.B. Maggirwar, S. Dewhurst, H.A. 30 1108 Gelbard, H.E. Gendelman, Neuroprotective activities of sodium valproate in a murine model of 31 1109 human immunodeficiency virus-1 encephalitis, J Neurosci. 23 (2003) 9162-9170. 32 1110 https://doi.org/10.1523/JNEUROSCI.23-27-09162.2003. 33 1111 [79] H. Dou, J. Morehead, J. Bradley, S. Gorantla, B. Ellison, J. Kingsley, L.M. Smith, W. Chao, G. 34 1112 Bentsman, D.J. Volsky, H.E. Gendelman, Neuropathologic and neuroinflammatory activities of hiv-1- 35 1113 infected human astrocytes in murine brain, Glia. 54 (2006) 81-93. https://doi.org/10.1002/glia.20358. 36 1114 [80] A.C. Ferraz, L.L. Xavier, S. Hernandes, M. Sulzbach, G.G. Viola, J.A. Anselmo-Franci, M. 37 1115 Achaval, C. Da Cunha, Failure of estrogen to protect the substantia nigra pars compacta of female 38 1116 rats from lesion induced by 6-hydroxydopamine, Brain Res. 986 (2003) 200-205. 39 1117 https://doi.org/10.1016/s0006-8993(03)03198-6. 40 1118 [81] L.L. Xavier, G.G. Viola, A.C. Ferraz, C. Da Cunha, J.M. Deonizio, C.A. Netto, M. Achaval, A 41 1119 simple and fast densitometric method for the analysis of tyrosine hydroxylase immunoreactivity in the 42 43 1120 substantia nigra pars compacta and in the ventral tegmental area, Brain Res Brain Res Protoc. 16 44 1121 (2005) 58-64. https://doi.org/10.1016/j.brainresprot.2005.10.002. 45 1122 [82] F.G. Martinez, E.E. Hermel, L.L. Xavier, G.G. Viola, J. Riboldi, A.A. Rasia-Filho, M. Achaval, 46 1123 Gonadal hormone regulation of glial fibrillary acidic protein immunoreactivity in the medial amygdala 47 1124 subnuclei across the estrous cycle and in castrated and treated female rats, Brain Res. 1108 (2006) 48 1125 117-126. https://doi.org/10.1016/j.brainres.2006.06.014. 49 1126 [83] L. Saur, P.P. Baptista, P.N. de Senna, M.F. Paim, P. do Nascimento, J. Ilha, P.B. Bagatini, M. 50 1127 Achaval, L.L. Xavier, Physical exercise increases gfap expression and induces morphological 51 1128 changes in hippocampal astrocytes, Brain Struct Funct. 219 (2014) 293-302. 52 1129 https://doi.org/10.1007/s00429-012-0500-8. 53 54 1130 55 56 57 1131 58 59 1132 60 61 27 62 63 64 65 1133 Figures 1 2 1134 Figures are provided as an attached single pdf document. 3 4 5 1135 6 7 1136 Figure Captions 8 1137 Figure 1: Experimental design scheme and TEM of the different nanomaterials tested. (A) 9 10 1138 Experimental design of the present study. After stereotactic administration of different nanomaterials, 11 12 1139 brains were collected at different time points. Molecular and cellular analyses were performed in the 13 1140 injection site and in nearby regions (anterior and posterior). (B) Transmission electron microscopy 14 15 1141 characterisation of the nanomaterials (aminated MWNTs, carboxylated MWNTs, GO, cationic and 16 1142 anionic liposomes) used in this work. 17 18 1143 19 1144 Figure 2: RT-qPCR analysis results obtained in the brain striatum injection site. (A) Gene 20 21 1145 expression levels of transcripts encoding pro-inflammatory cytokines, chemokines, and anti- 22 1146 inflammatory cytokines. The analysis was performed at 1 d, 2 d, and 7 d after injection of LPS, 5% 23 24 1147 dextrose, cationic or anionic liposomes (1 μg/μl), MWNT-NH3+ or ox-MWNT (0.5 μg/μl), or GO (0.5 25 1148 μg/μl). (B) Heat map presenting the statistical analysis. All statistical differences are shown in 26 27 1149 heatmap colours comparing dextrose with all types of material injected. Mean ± SEM, *p<0.05, 28 1150 **p<0.01, ***p<0.001 ****p<0.0001 vs 5% dextrose. 29 30 1151 31 1152 Figure 3: Quantitative analysis of glial cell immunohistochemical staining performed in 32 33 1153 different brain regions; anterior to the injection site (A), the injection site (B) and a posterior site (C) 34 1154 nearby the injection site. Relative proportion of the area (mean per ROI) covered by microglial 35 36 1155 (CD11b-immunopositive) cells (i), or astrocytes (GFAP-immunopositive cells) (ii) and estimated 37 1156 number of neurons (NeuN-immunopositive cells) (iii) counted in ROI were performed in the site of 38 39 1157 injection and in the regions anterior and posterior at 2 d after administration of LPS, 5% dextrose 40 1158 (ipsilateral and contralateral sides), MWNT-NH3+or ox-MWNT or GO (0.5 and 1 μg/μl), or cationic and 41 42 1159 anionic liposomes (1 μg/μl). Mean ± SEM; p-values are in comparison to 5% dextrose samples. 43 44 1160 *p<0.05, **p<0.01, ****p<0.0001 vs 5% dextrose. Representative images of the different 45 1161 immunostainings are presented in Figures S5, S6, and S7. 46 47 1162 48 1163 Figure 4: Representative images of cleaved-caspase 3 immunohistochemical staining. The 49 50 1164 immunophenotyping of cleaved-caspase 3 positive elements (undergoing apoptosis) in the injection 51 1165 site (the striatum) and in the anterior region distant from the injection site and in the posterior region 52 53 1166 adjacent to the injection site obtained at 2 d after injection of LPS, 5% dextrose, GO, and cationic 54 1167 liposomes. Note that apoptotic elements were visible in the injection site of LPS and cationic 55 56 1168 liposomes, while apoptotic cells visible in the GO-injected brain were comparable to the vehicle 57 1169 control (5% dextrose). 58 59 60 61 28 62 63 64 65 A B Aminated multi-walled CNTs Carboxylated multi-walled CNTs Graphene Oxide (MWNT-NH3+) (Ox-MWNT) (GO) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Cationic Liposomes Anionic Liposomes 19 20 (DOTAP:Chol) (DOPG:Chol) 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Figure 1 60 61 62 63 64 65 A Tnf-α Il-12 Il-1β Ifn-γ Il-6 Il-4 Il-10 Tgf-β Cxcl10 Ccl2 1 2 3 4 5 6 7 8 B Carbon based materials Lipid based materials 9 1d 1d 1d 2d 2d 2d 7d 7d 7d 1d 1d 2d 2d 7d 7d 10 11 Tnf-α Statistical difference Upregulated Downregulated 12 Il-1β from 5% dextrose 13 Il-6 **** ; p < 0.0001 14 15 Il-10 *** ; p < 0.001 16 Ccl2 17 ** ; p < 0.01 18 Il-12 19 Ifn-γ * ; p < 0.05 20 Il-4 21 ns 22 Tgf-β 23 Cxcl10 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Figure 2 60 61 62 63 64 65 A i) ii) iii) Anterior ** **** * 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl B i) ii) iii) Injection site ** * ** *** 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl C i) ii) iii) **** Posterior * **** 1 2 3 4 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl 0.5 μg/μl 1 μg/μl 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Figure 3 60 61 62 63 64 65 Cat. LPS 5% dextrose GO Liposomes Cleaved caspase 3 Anterior Striatum 50 μm Cortex Injection Striatum Posterior Striatum 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Figure 4 60 61 62 63 64 65 1170 Table of Content 1 2 1171 3 4 1172 Various nanomaterials are been explored for the brain delivery of therapeutic or imaging compounds. 5 1173 After administration in the brain, Portioli et al. demonstrate in a pilot study that functionalised carbon 6 7 1174 based nanomaterials may present a better safety profile than highly charged liposomes. Graphene 8 1175 Oxide nanosheets could even alleviate the inflammation caused by their direct injection in the brain. 9 10 1176 11 12 1177 Corinne Portioli, Cyrill Bussy, Mariarosa Mazza, Neus Lozano, Dhifaf A. Jasim, Maurizio Prato, 13 1178 Alberto Bianco, Marina Bentivoglio and Kostas Kostarelos 14 15 1179 16 1180 Injection of Graphene Oxide Nanosheets in the Brain Does not Induce Acute Neurotoxicity 17 18 1181 and Counteracts the Acute Microglial Activation related to Surgery in a Pilot Study 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 1182 43 44 1183 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 29 62 63 64 65 1184 Supplementary Information 1 2 1185 A supplementary information document with supplementary figures is provided as an attached single 3 4 1186 pdf document. 5 6 1187 7 8 1188 Supplementary Figure Captions 9 10 1189 Table S1: Forward and reverse sequences of the primers used in the present study. 11 12 1190 13 1191 Table S2: Primary and secondary antibodies and solutions used for immunohistochemical 14 15 1192 procedures. 16 1193 17 18 1194 Figure S1: Material characteristics. The different material characteristics (length, amount of 19 1195 amination and oxidation, zeta-potential, lateral dimensions, polydispersity, and thickness) are 20 21 1196 presented for (A) f-MWNTs (MWNT-NH3+ and ox-MWNT), (*) this data was previously published in 22 1197 Bussy et al. Nanoscale 2016 (ref. 55 in this article), (B) GO sheets, and (C) liposomes (DOTAP:Chol 23 24 1198 and DOPG:Chol). 25 1199 26 27 1200 Figure S2: RT-qPCR analysis of gene expression levels in the brain region posterior to the 28 1201 injection site. (A) Gene expression levels of transcripts encoding pro-inflammatory cytokines, 29 30 1202 chemokines, and anti-inflammatory cytokines. The analysis was performed at 1 d, 2 d, and 7 d after 31 1203 injection of LPS, 5% dextrose, cationic or anionic liposomes (1 μg/μl), MWNT-NH3+or ox-MWNT (0.5 32 33 1204 μg/μl), or GO (0.5 μg/μl). (B) Heat map presenting the statistical analysis. All statistical differences 34 1205 are shown in heat map colours comparing dextrose with all types of material injected. Mean ± SEM, 35 36 1206 *p<0.05, **p<0.01, ***p<0.001 ****p<0.0001 vs 5% dextrose. 37 1207 38 39 1208 Figure S3: RT-qPCR analysis of gene expression levels in the brain region anterior to the 40 1209 injection site. (A) Relative gene expression levels of transcripts encoding pro-inflammatory and anti- 41 42 1210 inflammatory cytokines and chemokines obtained at 1 d and 2 d after injection of LPS, 5% dextrose, 43 44 1211 cationic or anionic liposomes (1 μg/μl). (B) Heat map presenting the statistical analysis. All statistical 45 1212 differences are shown in heat map colours comparing dextrose with all types of material injected. 46 47 1213 Mean ± SEM, *p<0.05, **p<0.01, ***p<0.001 ****p<0.0001 vs 5% dextrose. 48 1214 49 50 1215 Figure S4: CD11b immunohistochemical staining of microglia. (A) Representative images of 51 1216 microglial cells as visualised by CD11b immunoreactivity in the three analysed brain regions (site of 52 53 1217 injection, anterior brain region distant from the site of injection, and posterior brain region adjacent to 54 1218 the site of injection) at 2 d after the injection of LPS, 5% dextrose, MWNT-NH3+ or ox-MWNT or GO 55 56 1219 (0.5 and 1 μg/μl), or cationic or anionic liposomes (1 μg/μl). (B) Densitometric evaluation (in optical 57 1220 density, or OD units) of the CD11b immunosignal intensity is shown in the injection site at 2 d after 58 59 60 61 30 62 63 64 65 1221 the administration of LPS, 5% dextrose, GO (0.5 and 1 μg/μl), or cationic or anionic liposomes (1 1 1222 μg/μl). Mean ± SEM; the p-value is for the comparison with 5% dextrose samples: *p<0.05, **p<0.01. 2 3 1223 4 1224 Figure S5: GFAP immunohistochemical staining of astrocytes. (A) Representative images of 5 6 1225 astrocytes as visualised by GFAP immunoreactivity in the three analysed brain regions (site of 7 1226 injection, anterior brain region to site of injection, and posterior brain region to site of injection) at 2 d 8 9 1227 after the injection of LPS, 5% dextrose, MWNT-NH3+or ox-MWNT or GO (0.5 and 1 μg/μl), or cationic 10 11 1228 or anionic liposomes (1 μg/μl). (B) Estimated number of astrocytes (per ROI) counted stereologically 12 1229 at the site of injection and in the regions anterior and posterior is shown at 2 d after injection of LPS, 13 14 1230 5% dextrose, or cationic liposomes. Mean ± SEM; the p value is for to the comparison with 5% 15 1231 dextrose samples: ***p<0.001. 16 17 1232 18 1233 Figure S6: NeuN immunohistochemical staining of neurons. Representative images of neuronal 19 20 1234 cells as visualised by NeuN immunoreactivity in the three analysed brain regions (site of injection, 21 1235 anterior brain region to site of injection and posterior brain region to site of injection) at 2 d after the 22 23 1236 injection of LPS, 5% dextrose, MWNT-NH3+or ox-MWNT or GO (0.5 and 1 μg/μl), or cationic or 24 1237 anionic liposomes (1 μg/μl). Quantitative evaluation of this staining is presented in Figure 3. 25 26 1238 27 1239 Figure S7: FluoroJade B staining. Representative images of Fluoro-Jade B staining in the cerebral 28 29 1240 cortex and in the striatum regions of the coronal brain sections containing the injection site at 2 d 30 1241 after administration of LPS, 5% dextrose, GO, or cationic liposomes. Degenerating neurons, labelled 31 32 1242 by green Fluoro-Jade B fluorescent staining, were visible mainly in the LPS- (cortex around the 33 1243 needle track and striatum) and cationic liposome (striatum)-treated tissues. 34 35 1244 36 1245 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 31 62 63 64 65 Copyright WILEY-VCH Verlag GmbH & Co. KGaA, 69469 Weinheim, Germany, 2018. 1 2 3 4 5 6 7 Supplementary Information 8 9 (Tables and Figures) 10 11 12 13 14 15 16 17 18 Injection of Graphene Oxide Nanosheets in the Brain Does not 19 20 Induce Acute Neurotoxicity and Counteracts the Acute Microglial 21 22 Activation related to Surgery in a Pilot Study 23 24 25 26 Corinne Portioli1, 2, $, Cyrill Bussy1, 3, 4 *, Mariarosa Mazza1, Neus Lozano1, 3, Dhifaf A. Jasim1,3, 27 Maurizio Prato5,6, Alberto Bianco7, Marina Bentivoglio2 and Kostas Kostarelos1,3,8, *, & 28 29 30 1 Nanomedicine Lab, School of Health Sciences, Faculty of Biology, Medicine & Health, The University of 31 Manchester, Manchester Academic Health Science Centre, Manchester, UK 32 2 Department 33 of Neurosciences, Biomedicine and Movement Sciences, University of Verona, Verona, Italy 34 3 National Graphene Institute, The University of Manchester, Manchester, UK 35 4 36 Lydia Becker Institute of Immunology and Inflammation, Faculty of Biology, Medicine and Health, The University 37 of Manchester, Manchester Academic Health Science Centre, Manchester, UK 38 5 Department of Chemical and Pharmaceutical Sciences, University of Trieste, 34127 Trieste, Italy 39 40 6 Carbon Nanobiotechnology Laboratory, CIC BiomaGUNE, 20009 San Sebastian, Spain. 41 7 42 University of Strasbourg, CNRS, Immunology, Immunopathology and Therapeutic Chemistry, UPR 3572, 43 Strasbourg, France 44 8 Catalan Institute of Nanoscience and Nanotechnology (ICN2), and The Barcelona Institute of Science and 45 Technology (BIST), Campus UAB, Bellaterra, Barcelona, Spain 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 1 63 64 65 Table S1: Forward and reverse sequences of the primers used in the present study. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Table S2: Primary and secondary antibodies and solutions used for immunohistochemical 28 procedures. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 2 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Figure S1: Material characteristics. The different material characteristics (length, amount of 44 amination and oxidation, zeta-potential, lateral dimensions, polydispersity, and thickness) are 45 presented for (A) f-MWNTs (MWNT-NH3+ and ox-MWNT), (*) this data was previously published in 46 47 Bussy et al. Nanoscale 2016 (ref. 55 in this article), (B) GO sheets, and (C) liposomes (DOTAP:Chol 48 and DOPG:Chol). 49 50 51 52 53 54 55 56 57 58 59 60 61 62 3 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Figure S2: RT-qPCR analysis of gene expression levels in the brain region posterior to the 53 injection site. (A) Gene expression levels of transcripts encoding pro-inflammatory cytokines, 54 chemokines, and anti-inflammatory cytokines. The analysis was performed at 1 d, 2 d, and 7 d after 55 injection of LPS, 5% dextrose, cationic or anionic liposomes (1 μg/μl), MWNT-NH3+or ox-MWNT (0.5 56 μg/μl), or GO (0.5 μg/μl). (B) Heat map presenting the statistical analysis. All statistical differences are 57 shown in heat map colours comparing dextrose with all types of material injected. Mean ± SEM, 58 *p<0.05, **p<0.01, ***p<0.001 ****p<0.0001 vs 5% dextrose. 59 60 61 62 4 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 Figure S3: RT-qPCR analysis of gene expression levels in the brain region anterior to the 55 injection site. (A) Relative gene expression levels of transcripts encoding pro-inflammatory and anti- 56 inflammatory cytokines and chemokines obtained at 1 d and 2 d after injection of LPS, 5% dextrose, 57 cationic or anionic liposomes (1 μg/μl). (B) Heat map presenting the statistical analysis. All statistical 58 differences are shown in heat map colours comparing dextrose with all types of material injected. 59 Mean ± SEM, *p<0.05, **p<0.01, ***p<0.001 ****p<0.0001 vs 5% dextrose. 60 61 62 5 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 Figure S4: CD11b immunohistochemical staining of microglia. (A) Representative images of 52 microglial cells as visualised by CD11b immunoreactivity in the three analysed brain regions (site of 53 injection, anterior brain region distant from the site of injection, and posterior brain region adjacent to 54 the site of injection) at 2 d after the injection of LPS, 5% dextrose, MWNT-NH3+ or ox-MWNT or GO 55 (0.5 and 1 μg/μl), or cationic or anionic liposomes (1 μg/μl). (B) Densitometric evaluation (in optical 56 density, or OD units) of the CD11b immunosignal intensity is shown in the injection site at 2 d after the 57 administration of LPS, 5% dextrose, GO (0.5 and 1 μg/μl), or cationic or anionic liposomes (1 μg/μl). 58 Mean ± SEM; the p-value is for the comparison with 5% dextrose samples: *p<0.05, **p<0.01. 59 60 61 62 6 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 Figure S5: GFAP immunohistochemical staining of astrocytes. (A) Representative images of 52 astrocytes as visualised by GFAP immunoreactivity in the three analysed brain regions (site of 53 injection, anterior brain region to site of injection, and posterior brain region to site of injection) at 2 d 54 55 after the injection of LPS, 5% dextrose, MWNT-NH3+or ox-MWNT or GO (0.5 and 1 μg/μl), or cationic 56 or anionic liposomes (1 μg/μl). (B) Estimated number of astrocytes (per ROI) counted stereologically 57 at the site of injection and in the regions anterior and posterior is shown at 2 d after injection of LPS, 58 5% dextrose, or cationic liposomes. Mean ± SEM; the p value is for to the comparison with 5% 59 dextrose samples: ***p<0.001. 60 61 62 7 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 Figure S6: NeuN immunohistochemical staining of neurons. Representative images of neuronal 44 cells as visualised by NeuN immunoreactivity in the three analysed brain regions (site of injection, 45 anterior brain region to site of injection and posterior brain region to site of injection) at 2 d after the 46 injection of LPS, 5% dextrose, MWNT-NH3+or ox-MWNT or GO (0.5 and 1 μg/μl), or cationic or 47 anionic liposomes (1 μg/μl). Quantitative evaluation of this staining is presented in Figure 3. 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 8 63 64 65 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Figure S7: FluoroJade B staining. Representative images of Fluoro-Jade B staining in the cerebral 18 cortex and in the striatum regions of the coronal brain sections containing the injection site at 2 d after 19 20 administration of LPS, 5% dextrose, GO, or cationic liposomes. Degenerating neurons, labelled by 21 green Fluoro-Jade B fluorescent staining, were visible mainly in the LPS- (cortex around the needle 22 track and striatum) and cationic liposome (striatum)-treated tissues. 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 9 63 64 65 Supporting Information Click here to access/download Supporting Information Portioli et al_SMALL_supp info_07092020.docx all figures as pdf format Click here to access/download Production Data Portioli et al _all figures included_SMALL_prod file.pdf supporting information file as word doc Click here to access/download Production Data Portioli et al_supp info_SMALL_07092020.docx main figures as pdf format Click here to access/download Production Data Portioli_et al _all main figures_SMALL_07092020.pdf main text word clean Click here to access/download Production Data Portioli et al_text _SMALL_09102020 R2_clean.docx