Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 www.chemeurj.org Graphene Oxide Nanosheets Hamper Glutamate Mediated Excitotoxicity and Protect Neuronal Survival In An In vitro Stroke Model Lorenza Tortella+,[a] Irene Santini+,[a] Neus Lozano,[b] Kostas Kostarelos,[b, c] Giada Cellot,*[a] and Laura Ballerini*[a] Dedicated to Professor Maurizio Prato on the occasion of his retirement. Small graphene oxide (s-GO) nanosheets reversibly downregu- show that the delivery of s-GO following OGD, during the late central nervous system (CNS) excitatory synapses, with endogenous build-up of secondary damage and excitotoxicity, potential developments as future therapeutic tools to treat improved neuronal survival. In a different paradigm, excitotox- neuro-disorders characterized by altered glutamatergic trans- icity cell damage was reproduced through exogenous gluta- mission. Excitotoxicity, namely cell death triggered by exceed- mate application, and s-GO co-treatment protected neuronal ing ambient glutamate fueling over-activation of excitatory integrity, potentially by directly downregulating the synaptic synapses, is a pathogenic mechanism shared by several neural over-activation brought about by exogenous glutamate. This diseases, from ischemic stroke to neurodegenerative disorders. proof-of-concept study suggests that s-GO may find novel In this work, CNS cultures were exposed to oxygen-glucose applications in therapeutic developments for treating excitotox- deprivation (OGD) to mimic ischemic stroke in vitro, and it is icity-driven neural cell death. Introduction One of the most promising GRMs are graphene oxide nano- sheets with small lateral size (< 500 nm, s-GO). This nano- During the last decade, graphene related materials (GRMs) were material has been reported both in vitro and in vivo to modulate increasingly exploited in biomedical applications[1–3] to promote specifically and selectively excitatory synapses,[10–12] by interfer- alternative therapeutic approaches and to cure brain ing with the presynaptic release of glutamate, leading to disorders.[4,5] More specifically, the reduced dimensions of GRMs, downregulation of glutamatergic synaptic transmission with no matching those of the smallest CNS functional units, such as impact on neuronal viability.[12,13] Recently, this ability was the sub-cellular structures composing synapses,[6–8] prompted applied to a rat model of anxiety disorder, in which symptoms their use to manipulate neuronal function during brain arose from a pathological glutamatergic over-signaling in the pathologies.[9] amygdala. A single intrathecal injection of s-GO delivered in the rodent during the disease consolidation normalized the [a] L. Tortella,+ I. Santini,+ Dr. G. Cellot, Prof. L. Ballerini amygdala glutamatergic transmission and anxiety related Neuroscience Area behaviors.[14] International School for Advanced Studies (SISSA/ISAS) Although this study supported the s-GO therapeutic value Via Bonomea 265, 34136 Trieste (Italy) in a specific disorder, it is still underexplored whether s-GO E-mail: cellot@sissa.it laura.ballerini@sissa.it could find a large-scale application for treating other neuro- [b] Dr. N. Lozano, Prof. K. Kostarelos logical diseases characterized by aberrant glutamatergic signal- Catalan Institute of Nanoscience and Nanotechnology (ICN2) ing. Since glutamate mediated excitotoxicity is a feature shared CSIC and BIST by the pathogenesis of various brain maladies, ranging from Campus UAB, Bellaterra, 08193 Barcelona (Spain) ischemic stroke or brain injury to the degenerative Parkinson’s [c] Prof. K. Kostarelos and Alzheimer’s diseases,[15] it is relevant to explore the use of s- Nanomedicine Lab, and Faculty of Biology, Medicine & Health, The National Graphene Institute GO to hinder excitotoxicity, thus preventing its downstream University of Manchester pathological outcomes. Manchester, M13 9PL (United Kingdom) In this work, we investigated the impact of s-GO treatments [+] These authors contributed equally to this work. on excitotoxicity by using an in vitro model of ischemic stroke. Supporting information for this article is available on the WWW under This is a pathological condition where brain OGD, typically https://doi.org/10.1002/chem.202301762 originating from the blockage of a brain artery,[16] due to energy This article is part of a joint Special Collection in honor of Maurizio Prato. supply disruption and related membrane depolarization, causes © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH an exceeding release of glutamate from pre-synaptic GmbH. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and re- terminals,[17] that induces, as secondary damage, a glutamate production in any medium, provided the original work is properly cited. Chem. Eur. J. 2023, 29, e202301762 (1 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 mediated excitotoxicity leading to neuroinflammation[18] and 2 layers) with a lateral dimension ranged from 100 to 500 nm augmenting cell death.[19] (Figure 1). By exposing dissociated CNS cells to in vitro OGD and To explore the potential activity by s-GO in counteracting assessing cell death through propidium iodide (PI) based assay brain excitotoxicity, we used an in vitro model of ischemic and immunofluorescence, we found that the application of s- stroke, that mimicked the lack of energy dependent glutamate- GO during the secondary damage preserved cell viability, induced excitotoxicity and the resulting cell death observed in increasing specifically neuronal survival. To better characterize the nervous tissue following acute ischemia in vivo.[23,24] To this the interaction between s-GO and neurons during excitotoxic aim, dissociated hippocampal cultures at the second week of secondary damage, we induced cell insult by direct adminis- differentiation in vitro were treated for 1 hour with an OGD tration of exogenous glutamate to cultures.[20] When s-GO was procedure consisting in the exposure to a medium deprived of co-applied with the exogenous neurotransmitter, cell death was glucose (OGD solution) in hypoxic atmosphere (see Experimen- prevented and loss in neuronal integrity largely decreased. tal section and Figure 2A). We selected a duration of the Supported also by our electrophysiological experiments, we treatment which, according to the literature, enables reproduc- propose that s-GO, through a downregulation of the exceeding ing a mild cell damage.[25,26] Control cultures were not exposed glutamatergic transmission, effectively protect neurons. Our to changes in glucose and oxygen concentrations. Immediately results highlight the role of s-GO in blocking upstream after the insult, the OGD solution was replaced with culturing excitotoxicity, thus extending the therapeutic applicability of medium and cultures were analyzed at different time points. the nanomaterial to the treatment of several neuro-disorders Due to energy unbalance, OGD leads first to a depolariza- characterized by aberrant glutamatergic signaling. tion of neuronal cell membrane[24,27] that in turn induces a massive release of glutamate neurotransmitter from the presynaptic site of synapses.[28,29] The released glutamate binds Results and Discussion and persistently activates its receptors on postsynaptic mem- branes, depolarizing neurons and escalating the pathogenic s-GO reverts OGD-induced cell death in ischemic stroke cascade leading to cell damage and death.[30,31] model in vitro To demonstrate that in our in vitro model of ischemic stroke cell damage following OGD was dependent on exceeding The s-GO material used in this work was synthesized following glutamatergic signaling, we used ionotropic receptors (GlutRs) the protocol already reported[21] and was characterized exten- antagonists to prevent cell death. Cultures were incubated sively in our previous publications.[12,13,22] As reported in Fig- during OGD and for the following 24 hours with CNQX (10 μM), ure 1, the specific s-GO batch used here was 1–2 nm thick (1– kynurenic acid (500 μM) and AP-V (25 μM), to block glutamater- gic AMPA, kainate and NMDA glutamate receptors respectively, while cell death was quantified after 48 hours from OGD through a PI based assay (Figure 2 A, B). PI is a cell membrane impermeable nucleic acid intercalating fluorescent dye, which labels selectively the nuclei of cells with damaged membrane, namely necrotic cells,[32] while the membrane permeable nuclear marker NucBlue (Hoechst 33342) was exploited to mark all nuclei present in the culture. Figure 2B shows representative confocal images of cultures that were co-labelled with PI (in red) and NucBlue (in blue). The total number of NucBlue positive cells/field remained constant among the different experimental conditions (see the Exper- imental section). However, upon OGD, PI positive-damaged cells were increased, while remained similar to the control level when protected by GlutRs antagonists. The percentage of PI necrotic cells (calculated on the total number of NucBlue positive cells) was 27 � 1% for controls, 39 � 1% for OGD treated cultures and 28 � 2% for OGD + GlutRs antagonists treated ones (bar plot in Figure 2C) and differences were statistically significant (control vs OGD and OGD vs OGD + GlutRs antago- nists, P < 0.05). Thus, OGD (1 hour) mimicked a moderate ischemic stress with a mild cell death, entirely mediated by glutamate depend- ent excitotoxicity. s-GO was shown to inhibit pre-synaptic release of glutamate in the CNS,[12,13] thus we investigated s-GO Figure 1. Structural and morphological characterization of s-GO nanosheets. (A) Height atomic force microscopy image (dimension: 5×5). (B) Scanning ability to hamper the exceeding glutamate release and to electron microscopy micrograph. reduce the OGD excitotoxicity. To this aim, we applied the Chem. Eur. J. 2023, 29, e202301762 (2 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 Figure 2. s-GO exerts a protective effect against secondary excitotoxic cell death in an in vitro model of ischemic stroke. (A) Schematic representation of the experimental protocol. To induce ischemia in hippocampal cultures, culturing medium (depicted in pink) was replaced with an OGD solution deprived of glucose (in white), and cultures were exposed in a hypoxic chamber to 95 % N2 and 5 % CO2 for 1 hours (red line). After the treatment, cultures were perfused with the initial culturing medium and re-exposed to an O2 enriched atmosphere. GlutRs antagonists (10 μM of CNQX, 500 μM of kynurenic acid and 25 μM of AP-V) were applied during OGD (in OGD solution, for 1 hour) and after it (in culturing medium, for 24 hours, light blue line). In a different set of samples, s-GO (10 μg/mL) was administered after OGD (in culturing medium, for 24 hours, green line). 48 hours after OGD, cultures were undergone to PI-based cell death assay and immunofluorescence. (B) Magnifications of representative fluorescence microscopy images obtained from cultures (untreated control, OGD, OGD + GlutRs antagonists and OGD + s-GO) undergone to PI based assay. In blue and red cell nuclei stained for NucBlue and PI, respectively. Note that the increment in cell death observed upon OGD was reverted by either GlutRs antagonists or s-GO applications. Scale bar 50 μm. (C) Bar plots showing the percentage of necrotic cells in the different conditions. Data are reported as mean � SEM, while dots superimposed to the bars correspond to the single field values. N = 63 fields for control and OGD cultures each, N = 26 fields for OGD + GlutR antagonists treated samples and N = 40 fields for OGD + s-GO treated samples, *P < 0.05. nanomaterial at a concentration of 10 μg/mL immediately after 48 hours after OGD, in the presence or absence of s-GO OGD and for the following 24 hours (in culturing medium). The (24 hours; sketched in Figure 2A), cultures were fixed and timeline of this application was chosen to mimic a possible immuno-stained with antibodies against the post mitotic neuro- therapeutic intervention, following ischemic stroke. The concen- nal marker NeuN[34] (in grey, Figure 3A) or with the glial fibrillary tration of s-GO used was in the range of that reported in acidic protein (GFAP, in green, Figure 3B), labelling astrocytic literature for inducing modulatory synaptic effects without cytoskeleton,[35] while nuclei were stained with DAPI[36] (in blue). affecting cell survival in dissociated cultures.[10,14] In s-GO treated Immunofluorescence microscopy revealed that cells dying due OGD cultures, we observed a statistically significant reduction to OGD were neurons, as their number decreased in a statisti- in the amounts of necrotic cells (31 � 2%, P < 0.05) respect to cally significant manner (P < 0.05) from 24 � 1 neurons/field in that of the untreated OGD cultures (Figure 2 B,C). control to 19 � 1 neurons/field in OGD treated samples (bar plot Similarly to GlutRs antagonists treatment, the application of in Figure 3C). No changes in glial cells densities were observed s-GO exerted a protective effect, decreasing the amount of between the two conditions (16 � 1 glial cells/field in control dying cells. and 16 � 1 glial cells/field in OGD treated samples; bar plot in Figure 3D; P > 0.05), indicating the lack of glial death or proliferation due to neuroinflammation.[37] s-GO prevents the In OGD cultures s-GO promotes neuronal survival without OGD reduction in healthy neurons (24 � 2 neurons/field), which affecting glial cells remained similar to that of control cultures (Figure 3C, OGD vs OGD + s-GO, P < 0.05), while the amount of glial cells was not After in vivo ischemic stroke neurons are the most vulnerable modified (14 � 1 glial cells/field in OGD + s-GO treated samples, phenotypes to OGD damage, while glial involvement is Figure 3D, P > 0.05). When measured in terms of GFAP intensity, observed only at a later stage.[33] We investigated the nature of indicative of glial over-reactivity,[38] all cultures displayed similar the OGD necrotic cells, to identify which cellular phenotype was values (403 � 8 arbitrary unit, a.u., in control, 434 � 14 a.u. in preserved by the s-GO treatment. Chem. Eur. J. 2023, 29, e202301762 (3 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 Figure 3. s-GO protects selectively neuronal survival after in vitro ischemia. (A-B) Magnifications of representative confocal microscopy images obtained from cultures in control condition and exposed to OGD with or without s-GO post-incubation (10 μg/mL), immunostained for the neuronal and glial markers NeuN (in grey, A) or GFAP (in green, B). In blue cell nuclei staining for DAPI. Note that OGD induced a decrease in the number of neurons that was restored by s-GO treatment, while glial cells were unaffected. Scale bar 50 μm. Bar plots showing the number of neurons (C), the number of GFAP-positive glial cells (D) and the mean GFAP fluorescence intensity (E). Data are reported as mean � SEM, while dots superimposed to the bars correspond to the single field values. For NeuN staining, N = 44 fields for control, N = 48 fields for OGD treated samples, N = 33 fields for OGD + s-GO treated samples; for GFAP staining, N = 18 fields for each condition; *P < 0.05. OGD treated samples, 500 � 37 a.u. in OGD + s-GO treated the processes undergoing during the secondary damage to samples; Figure 3E, P > 0.05). OGD. This analysis strongly suggested that, at least at the In a group of glutamate (30 min) treated cultures we co- considered time point (48 hours), in OGD cultures no patho- applied s-GO (10 μg/mL). After 30 minutes of incubation, the logical hallmarks could be detected in this cell population. culturing medium was replaced with a fresh one and, 24 hours later, cell death was evaluated through PI based assay (Fig- ure 4A). Control cultures were exposed neither to glutamate, s-GO protects hippocampal cells from necrosis induced by nor to s-GO. In Figure 4B, PI (in red) and NucBlue (in blue) co- exogenous glutamate application labeled cultures are shown with an enhanced percentage of necrotic cells (28 � 3 %) in glutamate treated samples when To support the glutamatergic pathway and excitotoxicity as the compared to controls (14 � 1 %), an effect prevented by s-GO pathogenic mechanisms targeted by s-GO in OGD, we used a (16 � 1 %; Figure 4C). The differences were statistically signifi- simplified in vitro model of glutamate-induced excitotoxicity, cant (for control vs glutamate and glutamate vs glutamate + where dissociated hippocampal cultures were treated directly sGO, P < 0.05). We measured no differences in NucBlue positive for 30 minutes with a high concentration of glutamate (40 μM cells densities among all treatments (see Experimental section). in culturing medium).[20] The application of the exogenous In addition, when cultures were analyzed for mitochondria neurotransmitter activates GlutRs, inducing cell depolarization damage through MitoTracker Red CM–XRos labelling, a similar and the release of further, endogenous, glutamate, responsible protective effect of the nanomaterial was detected (Supplemen- for a pathogenic loop sustaining excitotoxicity,[39,40] that mimics tary Figure S1). Chem. Eur. J. 2023, 29, e202301762 (4 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 Figure 4. s-GO prevents cell death upon primary excitotoxic insult induced by glutamate application. (A) Schematic representation of the experimental protocol. To induce direct excitotoxicity in hippocampal cultures, culturing medium was added with glutamate (40 μM, red line) for 30 minutes, with or without s-GO (10 μg/mL, green line). 24 hours after the excitotoxic insult, cultures underwent to PI-based cell death assay, electrophysiological recording to assess synaptic functionality and immunofluorescence to measure cell integrity. (B) Representative magnifications of confocal microscopy images in control, glutamate-treated and glutamate + s-GO treated cultures, co-stained for PI (in red) and NucBlue (in blue). Note that the glutamate induced increase in cell death was reverted by s-GO co-treatment. Scale bar 50 μm. (C) Bar plots showing the percentage of necrotic cells in the different conditions. Data are reported as mean � SEM, while dots superimposed to the bars correspond to the single field values. N = 18 fields for control, N = 12 fields for glutamate and glutamate + s-GO treated samples each, *P < 0.05. These results showed that the exogenous glutamate frequency depends on neuronal excitability and on the synaptic exposure triggered excitotoxicity and, as in OGD, s-GO network size (e. g. smaller networks show less activity[41]) and protected brain cells from death. sPSCs analysis allows to assess whether synapses supported efficiently inter neuronal communication via neurotransmitter release.[42] We observed in glutamate treated cultures a statisti- s-GO protects neuronal survival and neurites integrity during cally significant (P < 0.05) reduction in sPSCs frequency respect excitotoxicity to the untreated control (from 7 � 1.1 Hz to 2.8 � 0.8 Hz), with no changes in sPSCs amplitude (84.6 � 9.6 pA and 77 � 8.6 pA, In the next set of experiments, we used the glutamate-induced control and glutamate, respectively; Figure 5B,C). Upon induc- excitotoxicity paradigm to further address s-GO protective tion of excitotoxicity, the loss of neurons by induced cell death activity. s-GO was reported to modulate glutamate release from (see Figure 4B,C), results in a neuronal network of smaller size, presynaptic terminals[10,12] leading to down regulation of namely composed by a reduced number of cells respect to synaptic activity.[10,13,14] Based on these reports, we hypothesized control. In cultures co-treated with glutamate and s-GO, we that, s-GO, by inhibiting glutamate release and synaptic activity observed a similar reduction in sPSCs frequency (2.4 � 0.5 Hz, de-potentiated the excitotoxicity escalation, interrupting the control vs glutamate + s-GO), with no changes in sPSCs vicious circle, which fuels excitotoxic damage up to neuron amplitude (71.5 � 9.7 pA; Figure 5B,C). However, when cultures death. We used single cell patch clamp recordings to monitor were exposed to glutamate together with s-GO no increased spontaneous synaptic activity after glutamate insult, in cultures cell death was observed (see Figure 4B,C). Since the size of co-treated or not with s-GO (10 μg/mL; Figure 4A). control and glutamate + s-GO networks were similar, given that Figure 5A depicts exemplificative current recordings (volt- s-GO does not affect neuronal excitability,[14] the observed age clamp mode; 24 hours after treatments). The recorded decrease in synaptic activity should reflect a downregulation of neurons (with comparable membrane passive properties, see synaptic transmission mediated by s-GO[10,12,13,14] more than a Experimental section) displayed spontaneous postsynaptic reduction in neuronal firing activity. currents (sPSCs) activated by the release of neurotransmitters In a separated set of experiments, we applied s-GO (10 μg/ from the surrounding neuronal network, mostly of glutamater- mL) for 30 minutes to control cultures. Similarly to cultures co- gic nature in our cultures (see Experimental section). sPSCs incubated with glutamate + s-GO, when recording spontaneous Chem. Eur. J. 2023, 29, e202301762 (5 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 Figure 5. In a model of primary excitotoxicity, s-GO favors neuronal survival and integrity by interfering with synaptic transmission. (A) Representative traces from control (in black), glutamate (in red) and glutamate + s-GO treated cells (in green). Bar plots of sPSCs frequency (B) and amplitude (C). N = 16 in control, N = 15 in glutamate and N = 18 in glutamate + s-GO treated samples. Data are reported as mean � SEM, while dots superimposed to the bars correspond to single cell values. Magnifications of representative confocal microscopy images obtained from cultures in control condition and exposed to glutamate with or without s-GO (10 μg/mL), immunostained for the neuronal and glial markers β-tubulin III (in red, D) or GFAP (in green, E). In blue cell nuclei staining for DAPI. Note that glutamate application induced a loss in neuronal integrity and glial over-reactivity that were restored by s-GO treatment. Scale bar 50 μm. Bar plots showing the number of neurons with intact neurites (F), the number of GFAP-positive glial cells (G) and the mean GFAP fluorescence intensity (H). Data are reported as mean � SEM, while dots superimposed to the bars correspond to the single field values. For β-tubulin III staining, N = 22 in control and glutamate + s-GO treated samples and N = 21 for glutamate treated ones, for GFAP staining. *P < 0.05. activity after 24 hours from s-GO treatment, we observed a ure S2A,B), with no changes in sPSCs amplitudes (81 � 9 pA and statistically significant decrease (P < 0.05) in the frequency of 67 � 7 pA, control and s-GO respectively, Supplementary Fig- sPSCs respect to that of the untreated controls (from 0.76 � ure S2C), or in neuronal and glial densities or glial reactivity (P > 0.22 Hz to 0.28 � 0.14 Hz, control vs s-GO; Supplementary Fig- 0.05; Supplementary Figure S2D,E,F,G,H). This experiment, in- Chem. Eur. J. 2023, 29, e202301762 (6 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 dicating that in control s-GO downregulated synaptic activity as which the s-GO was co-applied. This model allowed to infer the expected,[10,12,13,14] suggested that similarly, when glutamate- mechanisms through which s-GO, by downregulating the treated cultures were co-incubated with s-GO, the detected activity of glutamatergic synapses, counteracted excitotoxic reduction in synaptic activity was due to the nanoparticle damage. modulation, regardless the preserved network size. s-GO prevented the excitotoxic cell death and this protec- We further assessed, 24 hours after the treatments, cell tive action was supposedly mediated by s-GO downregulation morphology and integrity through immunostaining for the of the neurotransmitter release,[10,12,13,14] which is augmented by neuronal marker β-tubulin III.[43] In Figure 5D control neurons exogenous glutamate.[40] We propose that the decreased are characterized by multiple neurites extending from the soma synaptic activity following glutamate exposure emerges mostly and forming a tangled network. Conversely, upon excitotoxicity from the reduction in circuit size due to neurons death,[41] as a massive loss in neuronal integrity was observed, resulting in supported by the immunostainings, and the measured mem- only few cells displaying projecting neurites. In each imaged brane passive values rule out changes in the membrane field, we quantified the number of neurons presenting intact potential. s-GO reduced PSCs frequency may instead reflect, the neurites extending � 10 μm from the cell body. These were nanomaterial induced downregulation in glutamate-mediated decreased from 9 � 1 neurons/field in control to 5 � 1 neurons/ neurotransmission,[10,12,13,14] ultimately protecting neurons from field in glutamate treated samples (Figure 5F; P < 0.05). In escalating excitotoxicity damage. cultures co-treated with glutamate and s-GO, neurites were Excitotoxicity triggered by exogenous glutamate induced similar to controls, as the number of neurons/field with intact glial over-reactivity, while no glial alterations were detected in neurites was 9 � 1 (for glutamate vs glutamate + s-GO, P < 0.05, OGD. This might be due to the mild OGD paradigm used, but Figure 5F). we cannot exclude that the timeline of our histological analysis In parallel experiments, we evaluated also glial density and precluded the characterization of a later glial involvement in reactivity, by staining the cultures for the astrocyte marker the secondary excitotoxic cascade.[33] GFAP (Figure 5E). While we found no changes in GFAP-positive The GFAP intensity increase in response to excitotoxicity is a cell densities among the different conditions (8 � 1 glial cells/ hallmark of neuroinflammation.[45,46] The s-GO ability to down- field in control, 9 � 1 glial cells/field in glutamate treated regulate reactive resident glial cells could be a direct effect on samples and 9 � 1 cells/field in glutamate + s-GO treated ones, astrocytes, namely not mediated by the reduced excitotoxicity. Figure 5G), GFAP intensity was enhanced by glutamate treat- This would be in agreement with our recent study,[22] reporting ment from 70 � 9 a.u. in control to 123 � 19 a.u. in glutamate s-GO downregulation of neuroinflammatory activated astro- treated samples (Figure 5H). Additionally, s-GO co-application cytes with s-GO preventing neuronal damage via multiple reverted GFAP reactivity at control level (58 � 5 a.u. in mechanisms. glutamate + s-GO treated samples, Figure 5H; for control vs glutamate and glutamate vs glutamate + s-GO, P < 0.05). In sum, these experiments confirmed the s-GO protective Conclusions action, during excitotoxicity. As suggested by the electro- physiological measures, we hypothesize that s-GO, via down- Our findings indicate that, by targeting the exceeding presy- regulation of glutamate release,[10,12,13,14] prevents dysregulated naptic glutamate release characterizing excitotoxicity, s-GO glutamatergic activity and protect neuronal survival during might be used for the development of alternative therapeutic glutamate overload. strategies against excitotoxicity pathogenic mechanisms. s-GO modulation of glutamatergic synapses has been Although several studies supported the in vitro and in vivo reported both in vitro and in vivo,[10,12,13] including pathological biocompatibility of s-GO,[10–14,22,48-50] further investigations are conditions characterized by dysfunctional synaptic plasticity.[14] required to support the translation of the nanomaterial in In this work, we report that s-GO efficiently counteracted clinical research. These will include the assessment of s-GO long neuronal cell death which follows excitotoxicity in vitro. term effects on nervous tissue to evaluate additional impact on We first focused on ischemic stroke, since the pathogenic glial cells, the therapeutic potential of the nanomaterial on cascade leading to cell death includes the excitotoxic preclinical models of ischemic stroke and s-GO toxicological process.[16,17,19,44] In this framework, in vitro models of ischemia[25] assessment in course of neuropathologies. allow to dissect out in a simplified and accessible setting s-GO effects on neurons during excitotoxicity. In our model, excitotoxicity is secondary to OGD, thus Experimental Section mimicking closely the sequence of pathogenic events occurring Synthesis and characterization of s-GO. Biological-grade s-GO in vivo[25] and allowing to test the impact of s-GO on cell material was synthesized in house from graphite powder (Sigma damage. In this paradigm, we observed that the nanomaterial Aldrich, UK) by the modified Hummers method already described applied in a therapeutic fashion after OGD (i. e. after the stroke), under endotoxin-free conditions.[21,51] The specific batch for the s- could prevent cell damage, even more efficiently than the GO material used in this study was fully characterized and the detailed information for the characterization can be found in.[22] GlutRs antagonists cocktail. To focus on glutamate-mediated pathways, we reproduced Materials used for neurobiology experiments. All materials were excitotoxicity by exogenous glutamate exposure[20] during bought from Sigma-Aldrich company if not differently stated. Chem. Eur. J. 2023, 29, e202301762 (7 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 Preparation of rat dissociated hippocampal cultures. Dissociated with a Nikon C2 confocal microscope with a 20X objective (NA hippocampal cultures were prepared from 2 to 3 days postnatal 0.50). The analysis was performed offline using the image- (P2 P3) Wistar rats. All procedures were done in agreement with processing CellCounting plug-in of Fiji software. Necrotic cells were the Italian law (decree 26/14) and the EU guidelines (2007/ 526/CE quantified as percentage of total number of the NucBlue positive and 2010/63/UE). The animal use was authorized by the Italian cells present in each field. Ministry of Health (authorization number: 22DAB.NYQA) and approved by the local veterinary authorities. Once isolated, hippo- For the OGD model, two sets of experiments were carried out and campi were treated by enzymatic and mechanical digestion for each the number of cells positive to NucBlue were similar following a procedure describe elsewhere.[52] The obtained cells among the different conditions. In the first set of experiments, were seeded on poly-L-ornithine-coated glass coverslips these were 148 � 11 in control, 171 � 11 in OGD and 154 � 11 in (24×12 mm2, Kindler, EU) at a density of 150,000 cells/mL. Neuronal OGD + GlutRs antagonists treated samples (P > 0.05), while in cultures were maintained in stable conditions (37 °C, 5 % CO2) in second one these were 80 � 4 in control, 81 � 4 in OGD and 85 � 4 Neurobasal-A Medium (Gibco) supplemented with 10 mM Gluta- in OGD + s-GO treated samples (P > 0.05). MAX (Gibco), 2 % B27 supplement (Life Technologies, no. For the primary damage model, the total number of NucBlue 17504044), and 500 nM gentamycin and replaced after 3 days with positive cells/field was unchanged among the different experimen- fresh medium. Cultures were used for experiments in the second tal conditions (364 � 20 cells/field in untreated control, 358 � 24 week of differentiation in vitro. cells/field in glutamate treated samples and 379 � 29 cells/field in OGD treatment. Dissociated hippocampal cultures were exposed to glutamate + s-GO treated samples, p > 0.05). an OGD treatment, according to previous literature.[25] Briefly, Immunofluorescence labelling. Hippocampal cells were washed in neurobasal culturing medium was removed and replaced by 2 mL phosphate buffered saline (PBS) solution and then fixed with 4 % of sterile OGD medium, lacking glucose and composed as follows paraformaldehyde (PFA) for 20 min at room temperature (RT). Cells (in mM): 130 NaCl, 5.4 KCl, 1.8 CaCl2, 26 NaHCO3, 0.8 MgSO4 · 7H2O, were permeabilized with 3 % Triton X-100 and blocked with 5 % 1.18 NaH2PO4. Cell cultures were transferred to a hypoxic chamber, NGS in PBS for 60 min at RT, and then incubated in PBS with 5 % fabricated by SISSA mechatronic facility, equipped with a pressure FBS and primary antibodies for 60 min. The primary antibodies lid and an inlet and outlet tubes with respective valves and flushed used were mouse monoclonal anti-NeuN (Millipore, 1 : 500 dilution), by a gas mixture consisting of 95 % N2 and 5 % CO2 for 5 min at rabbit polyclonal anti-β-tubulin III (Sigma, 1 : 500 dilution), mouse 15 L/min. Next, the valves were closed and the chamber was placed polyclonal anti-GFAP (Sigma, 1 : 500 dilution). After the primary in a humidified incubator (37 °C) for 55 minutes. After OGD treat- incubation and PBS washes, cells were incubated for 60 min with ment (lasting overall 1 hour), cell cultures were taken from the the secondary antibodies Alexa Fluor 488 goat anti-mouse chamber and the OGD medium was replaced with the previous (Invitrogen, no. A-11001; dilution 1 : 500), Alexa Fluor 594 goat anti- culturing medium. The control cells did not undergo to any rabbit (Invitrogen, no. A-11008; dilution 1 : 500). DAPI (Invitrogen, medium change and were maintained in stable conditions (37 °C, no. D1306; dilution 1 : 500) was used to stain the nuclei. Samples 5 % CO2) for the same time of OGD exposure. were mounted in anti-fade medium Fluoromont-G™ (Invitrogen™ In cultures treated with s-GO after OGD, the nanomaterial was 00495802) on 24 x 60 mm2 glass slide (0.13-0.17 mm thick, Menzel- Gläser). Image acquisition was performed using a confocal micro- added to the culturing medium at the concentration of 10 μg/mL scope (Nikon C2-si Laser Scanning Confocal Microscope) with 20x for 24 hours. The selected concentration of nanomaterial was in the range of those reported in literature to exert effects on neuronal (0.50 NA) and 40x (0.75 NA) air objectives (Z-stacks were acquired activity without cytotoxicity in culture.[10,14] At the end of the every 500 nm). incubation, the medium was replaced with fresh one. Mitochondrial imaging. Cultures were treated with glutamate in Primary damage model. Under a biological hood, glutamate (L- the presence or absence of s-GO and with s-GO alone as Glutamic acid potassium salt monohydrate, stock solution 10 mM) supramentioned and compared with untreated (control) ones. Cells was added to culturing medium at the final concentration of were labelled with MitoTracker™ Red CM–XRos (2 μM, M7512; 40 μM. Then cultures were incubated at 37 °C for 30 minutes. Invitrogen; Thermo Fisher Scientific, Inc.), according to,[53] together with NucBlue dye, at 37°C in a humidified 5 % CO2 atmosphere for Glutamate concentration and incubation time were in the range of 24 hours. Then cells were rinsed twice with PBS, mounted on a those reported in literature to induce a significant increase in cell death after 24 hours from the exposure.[20] For s-GO treatments, glass slide and imaged at the confocal microscope with a 40X cultures were incubated with glutamate (40 μM) and s-GO (10 μg/ objective. The number of cells co-stained for MitoTracker Red CM– mL) simultaneously. At the end of the incubation time, the medium XRos and NucBlue was manually counted using Fiji software. was washed out and replaced with fresh culturing medium for the Electrophysiological recordings. Neuronal activity was recorded following 24 hours. from hippocampal cultures through single cell patch clamp PI based necrosis assay. PI is a highly polar (not lipidic permeant) technique. Voltage clamp whole-cell recordings were performed at fluorescent dye used to assess neuronal death.[25] Applied to RT with pipettes (3–5 MΩ) containing (in mM): 120 K gluconate, 20 KCl, 2 MgCl2, 10 EGTA, 10 HEPES, 2 Na2ATP, pH 7.3. The cultures together with the NucBlue dye, a cell-permeant nuclear extracellular solution was the one used for PI based assay. Cultures staining, can be exploited to calculate the percentage of necrotic cells in the preparation. Under a chemical hood, the staining were mounted on a chamber and visualized with an inverted solution was prepared adding PI at the final concentration of 1 μg/ microscope (Nikon Eclipse Ti-U). During experiments, neurons were mL and one drop of NucBlue (NucBlue™ Live ReadyProbes™ continuously perfused with an extracellular solution through a Reagent, R37605-Invitrogen™) to the extracellular physiological gravity-based perfusion system at a rate of 1 mL/min. Recordings solution composed (in mM) as follows: 150 NaCl, 4 KCl, 1 MgCl2, 2 were performed through a Multiclamp 700B patch amplifier (Axon CaCl2, 10 HEPES, 10 glucose, pH 7.4. Next, the culture medium was CNS, Molecular Devices) with a sampling rate of 10 kHz. Data were removed from the petri dishes and cells were washed four times acquired using a pClamp 10.3 software (Molecular Devices LLC, with the extracellular solution. Cells were treated with the staining USA). Capacitative transients due to the recording pipette were solution for 20 minutes. After mounting the samples on a glass compensated. All the recorded neurons were characterized by a slide (in staining solution), images (3 fields/coverslip) were acquired series resistance less than 1/10 of the membrane resistance and it Chem. Eur. J. 2023, 29, e202301762 (8 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 remained stable for the entire duration of the recording. In order to Fontanini for technical support for cell cultures and to Prof. measure the input resistance and cell capacitance, a stimulation Gabriele Baj for his support regarding mitochondria labelling. protocol was applied while the cell was clamped close to the resting membrane potential ( 56 mV). It consisted in 100 ms long lasting depolarizing steps of voltage (10 mV of amplitude), repeated 10 times. For single cells, the currents elicited by these stimulations Conflict of Interests were recorded and averaged. The input resistance was calculated according to Ohm’s law: ΔV=R ΔI (ΔV: applied pulse; R: resistance; The authors declare no conflict of interest. ΔI: current). The membrane capacitance was calculated as the integral in time of the capacitive transient (elicited by the voltage step), divided for the amplitude of the applied ΔV. Resting membrane potential was measured in current clamp in I = 0. These Data Availability Statement passive membrane properties, used as indicators of neuronal degree of maturation and health,[54,55,56] were analysed offline with The data that support the findings of this study are available the software Clampfit 10.6. from the corresponding author upon reasonable request. Upon the different treatments, no significant difference between the experimental groups were observed in membrane capacitance Keywords: graphene oxide · excitotoxicity · in vitro ischemic (control: 73 � 3 pF, glutamate: 62 � 3 pF, glutamate + s-GO: 63 � 4 stroke · nanomedicine · glutamatergic synapses pF), input resistance (control: 456 � 65 MΩ, glutamate: 587 � 47 MΩ, glutamate + s-GO: 533 � 56 MΩ) and resting membrane potential (control: 60 � 2, glutamate: 60 � 1, glutamate + s-GO: 55 � 2 mV). This indicated that recordings were performed from three groups of neurons presenting homogeneous properties. [1] Y. Yang, A. M. Asiri, Z. Tang, D. Du, Y. Lin, Mater. Today 2013, 16 (10), 365–373. Recordings of spontaneous postsynaptic activity were performed at [2] K. Kostarelos, K. S. Novoselov, Science 2014, 344 (6181), 261–263. a holding potential of 56 mV (not corrected for liquid junction [3] A-I. Lazăr, K. Aghasoleimani, A. Semertsidou, J. Vyas, A-L. Roşca, D. Ficai, potential, which was 14 mV). For each trace, amplitude and A. Ficai, Nanomaterials 2023, 13, 1092. frequency of sPSCs were evaluated offline with Axograph 1.4.4 [4] K. E. Kitko, Q. Zhang, Front. Neurol. Neurosci. 2019, 13, 26. event detection software. This exploits a detection algorithm based [5] M. K. Akbari, N. S. Lopa, M. Shahriari, A. Najafzadehkhoee, D. Galusek, S. on sliding templates to separate glutamatergic fast decaying Zhuiykov, J. Funct. Biomater. 2023, 14, 35. [6] B. Zhang, Y. H. Koh, R. B. Beckstead, V. Budnik, B. Ganetzky, H. J. Bellen, (~4 ms) and GABAergic slow decaying (~25 ms) postsynaptic Neuron 1998, 21, 1465–75. currents on the basis of their different decay times. In control [7] K. R. Smith, K. J. Kopeikina, J. M. Fawcett-Patel, K. Leaderbrand, R. Gao, cultures, 74 % of sPSCs were of glutamatergic nature. B. Schürmann, K. Myczek, J. Radulovic, G. T. Swanson, P. Penzes, Neuron 2014, 84, 399–415. Statistical analysis. All the statistical analyses were carried out [8] E. Pchitskaya, I. Bezprozvanny, Front. Synaptic Neurosci. 2020, 12, 31. using GraphPad Prism 8. The Gaussian distribution of data was [9] G. Cellot, A. Franceschi Biagioni, L. Ballerini, Pediatr. Res. 2022, 92, 71– assessed through Shapiro-Wilk test. If all the data compared in one 79. set of experiments resulted parametric, next statistically significant [10] R. Rauti, N. Lozano, V. León, D. Scaini, M. Musto, I. Rago, F. P. differences were assessed by one way ANOVA test, using Tukey’s Ulloa Severino, A. Fabbro, L. Casalis, E. Vázquez, K. Kostarelos, M. Prato, Multiple Comparisons as post hoc test. In opposition, if data were L. Ballerini, ACS Nano 2016, 10, 4459–71. not distributed as in a Gaussian distribution, non-parametric [11] N. Secomandi, A. Franceschi Biagioni, K. Kostarelos, G. Cellot, L. Ballerini, Nanomedicine 2020, 26, 102174. analysis was used (Kruskal-Wallis test was performed and Dunn’s [12] G. Cellot, S. Vranic, Y. Shin, R. Worsley, A. F. Rodrigues, C. Bussy, C. Multiple Comparison was used as post hoc test). p values were Casiraghi, K. Kostarelos, J. R. McDearmid, Nanoscale Horiz. 2020, 5, considered statistically significant when minor than 0.05. Data are 1250–1263. reported as mean with the standard error of the mean (SEM). N is [13] R. Rauti, M. Medelin, L. Newman, S. Vranic, G. Reina, A. Bianco, M. Prato, the number of cells or analyzed fields. K. Kostarelos, L. Ballerini, Nano Lett. 2019, 19, 2858–2870. [14] A. Franceschi Biagioni, G. Cellot, E. Pati, N. Lozano, B. Ballesteros, R. Casani, N. C. Coimbra, K. Kostarelos, L. Ballerini, Biomaterials 2021, 271, 120749. Acknowledgements [15] A. Granzotto, L. M. T. Canzoniero, S. L. Sensi, Front. Mol. Neurosci. 2020, 13, 600089. [16] B. Puig, S. Brenna, T. Magnus, Int. J. Mol. Sci. 2018, 19, 2834. This work received funding from the European Union Horizon [17] M. Szatkowski, D. Attwell, Trends Neurosci. 1994, 17, 359–365. 2020 Research and Innovation Programme under Grant Agree- [18] C. Iadecola, J. Anrather, Nat. Med. 2011, 17, 796. ment No. GrapheneCore3 (881603). We would like to acknowl- [19] S. G. Daniele, G. Trummer, K. A. Hossmann, Z. Vrselja, C. Benk, K. T. Gobeske, D. Damjanovic, D. Andrijevic, J. S. Pooth, D. Dellal, F. edge Dr. Michelle Ntola for the synthesis of the specific GO Beyersdorf, N. Sestan, Nat. Rev. Neurosci. 2021, 22, 553–572. batch used in this work and Dr. Luis Miguel Arellano and [20] A. P. Simões, C. G. Silva, J. M. Marques, D. Pochmann, L. O. Porciúncula, Angeliki Karakasidi for their contributions to the full character- S. Ferreira, J. P. Oses, R. O. Beleza, J. I. Real, A. Köfalvi, B. A. Bahr, J. Lerma, R. A. Cunha, R. J. Rodrigues, Cell Death Dis. 2018, 9, 297. ization of the s-GO material. The ICN2 has been supported by [21] A. F. Rodrigues, L. Newman, N. Lozano, S. P. Mukherjee, B. Fadeel, C. the Severo Ochoa Centres of Excellence programme [SEV-2017- Bussy, K. Kostarelos, 2D Mater. 2018, 5 (3), 035020. 0706] and is currently supported by the Severo Ochoa Centres [22] G. Di Mauro, R. Amoriello, N. Lozano, A. Carnasciali, D. Guasti, M. Becucci, G. Cellot, K. Kostarelos, C. Ballerini, L. Ballerini, ACS Nano 2023, of Excellence programme, Grant CEX2021-001214-S, both 17, 1965–1978. funded by MCIN/AEI/10.13039.501100011033. We are grateful [23] M. P. Goldberg, D. W. Choi, J. Neurosci. 1993, 13 (8), 3510–24. to Marco Gigante (SISSA Mechatronic lab) for the help in setting [24] R. Wang, X. Zhang, Y. Fan, Y. Shen, W. Hu, Z. Chen, PLoS One 2012, 7 (5), e37574. up the apparatus for OGD, to Raffaele Casani and Mario [25] C. I. Tasca, T. Dal-Cim, H. Cimarosti, Methods Mol. Biol. 2015, 1254, 197– 210. Chem. Eur. J. 2023, 29, e202301762 (9 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH 15213765, 2023, 67, Downloaded from https://chemistry-europe.onlinelibrary.wiley.com/doi/10.1002/chem.202301762 by Csic Organización Central Om (Oficialia Mayor) (Urici), Wiley Online Library on [17/01/2024]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Research Article Chemistry—A European Journal doi.org/10.1002/chem.202301762 [26] J. Zhang, L. Huang, X. Shi, L. Yang, F. Hua, J. Ma, W. Zhu, X. Liu, R. Xuan, [44] T. W. Lai, S. Zhang, Y. T. Wang, Prog. Neurobiol. 2014, 115, 157–88. Y. Shen, J. Liu, X. Lai, P. Yu, Aging (Albany NY) 2020, 12 (23), 24270– [45] P. M. George, G. K. Steinberg, Neuron 2015, 87 (2), 297–309. 24287. [46] M. Pekny, M. Pekna, Physiol. Rev. 2014, 94 (4), 1077–1098. [27] M. I. Cuartero, J. de la Parra, A. Garcia-Culebras, I. Ballesteros, I. Lizasoain, [47] H. P. Makarenkova, V. I. Shestopalov, Front. Plant Physiol. 2014, 5, 63. M. A. Moro, Curr. Pharm. Des. 2016, 22, 1060–1073. [48] D. A. Jasim, C. Ménard-Moyon, D. Bégin, A. Bianco, K. Kostarelos, Chem. [28] S. Y. Lee, J. H. Kim, J. Physiol. 2015, 593 (13), 2793–806. Sci. 2015, 6, 3952–3964. [29] A. L. Andrade, D. J. Rossi, J. Physiol. 2010, 588 (9), 1499–1514. [49] A. F. Rodrigues, L. Newman, D. Jasim, S. P. Mukherjee, J. Wang, I. A. [30] K. Kuboyama, H. Harada, H. Tozaki-Saitoh, M. Tsuda, K. Ushijma, K. Vacchi, C. Ménard-Moyon, A. Bianco, B. Fadeel, K. Kostarelos, C. Bussy, Inoue, J. Cereb. Blood Flow Metab. 2011, 31 (9), 1930–1941. Adv. Sci. 2020, 7 (12),1903200. [31] D. W. Choi, J. Neurobiol. 1992, 23 (9), 1261–76. [50] S. P. Mukherjee, A. R. Gliga, B. Lazzaretto, B. Brandner, M. Fielden, C. [32] L. C. Crowley, A. P. Scott, B. J. Marfell, J. A. Boughaba, G. Chojnowski, Vogt, L. Newman, A. F. Rodrigues, W. Shao, P. M. Fournier, M. S. Toprak, N. J. Waterhouse, Cold Spring Harb. Protoc. 2016, 2016 (7), 10.1101/ A. Star, K. Kostarelos, d K. Bhattacharyaa, B. Fadeel, Nanoscale 2018, 10, pdb.prot087163. 1180–1188. [33] Y. Du, W. Wang, A. D. Lutton, C. M. Kiyoshi, B. Ma, A. T. Taylor, J. W. [51] S. P. Mukherjee, N. Lozano, M. Kucki, A. E. Del Rio-Castillo, L. Newman, E. Olesik, D. M. McTigue, C. C. Askwith, M. Zhou, Exp. Neurol. 2018, 303, 1– Vázquez, K. Kostarelos, P. Wick, B. Fadeel, PLoS One 2016, 11 (11), 11. e0166816. [34] W. Duan, Y. P. Zhang, Z. Hou, C. Huang, H. Zhu, C. Q. Zhang, Q. Yin, Mol. [52] G. Cellot, F. M. Toma, Z. K. Varley, J. Laishram, A. Villari, M. Quintana, S. Neurobiol. 2016, 53, 1637–1647. Cipollone, M. Prato, L. Ballerini, J. Neurosci. 2011, 31, 12945–53. [35] E. M. Hol, M. Pekny, Curr. Opin. Cell Biol. 2015, 32, 121–130. [53] Y. Zhang, N. S. Gliyazova, P. A. Li, G. Ibeanu, Exp. Ther. Med. 2021, 21, [36] J. Kapuscinski, Biotech. Histochem. 1995, 70 (5), 220–233. 221. [37] C. Y. Li, X. Li, S. F. Liu, W. S. Qu, W. Wang, D. S. Tian, Neurochem. Int. [54] J. S. Carp, J. Neurophysiol. 1992, 68, 1121–1132. 2015, 83, 9–18. [55] Y. Gao, L. Liu, Q. Li, Y. Wang, Neurosci. Lett. 2015, 591, 138–143. [38] D. Schiffer, M. T. Giordana, A. Migheli, G. Giaccone, S. Pezzotta, A. [56] U. Djuric, A. Y. Cheung, W. Zhang, R. S. Mok, W. Lai, A. Piekna, J. A. Mauro, Brain Res. 1986, 374 (1), 110–118. Hendry, P. J. Ross, P. Pasceri, D. S. Kim, M. W. Salter, J. Ellis, Neurobiol. [39] Doble A, Pharmacol. Ther. 1999, 81, 163–221. Dis. 2015, 76, 37–45. [40] C. M. Norris, E. M. Blalock, O. Thibault, L. D. Brewer, G. V. Clodfelter, N. M. Porter, P. W. Landfield, J. Neurophysiol. 2006, 96 (5), 2488–500. [41] E. Cohen, M. Ivenshitz, V. Amor-Baroukh, V. Greenberger, M. Segal, Brain Res. 2008, 1235, 21–30. [42] J. Simonnet, L. Richevaux, D. Fricker, Methods Mol. Biol. 2021, 2188, 285– 309. Manuscript received: June 5, 2023 [43] C. D. Katsetos, A. Legido, E. Perentes, S. J. Mörk, J. Child Neurol. 2003, 18, Accepted manuscript online: September 14, 2023 851–66. Version of record online: October 20, 2023 Chem. Eur. J. 2023, 29, e202301762 (10 of 10) © 2023 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH