Biomaterials 271 (2021) 120749 Contents lists available at ScienceDirect Biomaterials journal homepage: www.elsevier.com/locate/biomaterials Graphene oxide prevents lateral amygdala dysfunctional synaptic plasticity and reverts long lasting anxiety behavior in rats Audrey Franceschi Biagioni a, 1, Giada Cellot a, 1, Elisa Pati a, Neus Lozano b, Belén Ballesteros b, Raffaele Casani a, Norberto Cysne Coimbra c, Kostas Kostarelos b, d, Laura Ballerini a, * a Neuron Physiology and Technology Lab, International School for Advanced Studies (SISSA), Neuroscience, Via Bonomea 265, 34136, Trieste, Italy b Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, Bellaterra, 08193, Barcelona, Spain c Laboratory of Neuroanatomy & Neuropsychobiology, Department of Pharmacology, Ribeirão Preto Medical School of the University of São Paulo (FMRP-USP), Av. Bandeirantes 3900, 14049-900, Ribeirão Preto, SP, Brazil d Nanomedicine Lab, National Graphene Institute and Faculty of Biology, Medicine & Health, The University of Manchester, AV Hill Building, Oxford Rd, Manchester, M13 9PL, United Kingdom A R T I C L E I N F O A B S T R A C T Keywords: Engineered small graphene oxide (s-GO) sheets were previously shown to reversibly down-regulate gluta­ Graphene-based nanovectors matergic synapses in the hippocampus of juvenile rats, disclosing an unexpected translational potential of these Biotechnology of brain disorders nanomaterials to target selective synapses in vivo. Synapses are anatomical specializations acting in the Central Synaptic plasticity Nervous System (CNS) as functional interfaces among neurons. Dynamic changes in synaptic function, named Patch-clamp Amygdala synaptic plasticity, are crucial to learning and memory. More recently, pathological mechanisms involving dysfunctional synaptic plasticity were implicated in several brain diseases, from dementia to anxiety disorders. Hyper-excitability of glutamatergic neurons in the lateral nucleus of the amygdala complex (LA) is substantially involved in the storage of aversive memory induced by stressful events enabling post-traumatic stress disorder (PTSD). Here we translated in PTSD animal model the ability of s-GO, when stereotaxically administered to hamper LA glutamatergic transmission and to prevent the behavioral response featured in long-term aversive memory. We propose that s-GO, by interference with glutamatergic plasticity, impair LA-dependent memory retrieval related to PTSD. 1. Introduction treat pathologies involving alterations in excitatory neurotransmission. New biotechnology-based therapeutic interventions for brain pa­ Emerging evidence in mammals indicates the amygdala complex as thologies have involved engineering of novel nanomaterials, such as the brain structure implicated in the development of anxiety disorder graphene, the one carbon atom thick 2-dimensional (2D) material [13]. [1–3]. In animal models, core behavior related to anxiety can be induced Graphene or its derivatives such as its oxide form, graphene oxide (GO), by responses to fear stimuli [4,5] and previous elegant experiments have been exploited as components of bioelectronic devices, as (nano) suggested that the over-activation of distinct populations of basal vectors in drug-delivery platforms and engineered as promising tissue amygdala neurons, such as glutamatergic ones effectively sustain fear scaffolds [14–17]. We have previously shown the ability of small gra­ behavior [6–8]. Moreover, dysfunctions in the glutamatergic system phene oxide nanosheets (s-GO), to transiently impair glutamatergic have a primary role in several brain diseases [9] and severe mood dis­ transmission in the rat hippocampus, both in vitro and in vivo, presum­ orders are increasingly associated to altered synaptic plasticity [10]. In ably with a direct interference with the presynaptic glutamate release such a context, the glutamatergic system, known to play a major role in machinery [18]. More recently, additional studies sustained s-GO syn­ tuning neuronal plasticity [11,12], may represent a promising target to apse specificity and kinetics when injected in specific Central Nervous Abbreviations: EPM, elevated plus maze; EPSC, glutamatergic postsynaptic currents; IPSC, inhibitory postsynaptic currents; LA, lateral amygdala; OF, open field apparatus; s-GO, small graphene oxide. * Corresponding author. E-mail address: laura.ballerini@sissa.it (L. Ballerini). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.biomaterials.2021.120749 Received 30 November 2020; Received in revised form 23 February 2021; Accepted 27 February 2021 Available online 4 March 2021 0142-9612/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 System (CNS) areas and further showed in zebrafish larvae in vivo that elevated plus maze (EPM). This apparatus consisted of four arms (50 × s-GO injected directly to the spinal cord were able to reduce gluta­ 10 × 40 cm), two open arms (without walls) and two closed arms (with matergic transmission, effectively impairing locomotor behavior [19]. 40 cm high walls) connected by a central square (10 × 10 cm). The maze In the current work, a rat model of an anxiety disorder (post-traumatic was elevated 50 cm from the ground. Rats (n = 6 per group) were placed stress disorder (PTSD)) was established, and a suspension of thin s-GO in the closed arm and were allowed to freely explore the apparatus for 5 nanosheets was administered in a single dose in the LA to prevent min. Duration and frequency of entrance in the open and closed zone pathological long-term enhancement in excitatory neurotransmission. was evaluated. Entry into a zone was scored when the center-point of the Behavioral studies combined with tissue histology, confocal microscopy rat’s body was within it. We adopted an anxiety index which was pre­ and ex vivo electrophysiology, demonstrated that s-GO sheets were able viously validated [24] and calculated for each rat based on the EPM to block LA glutamatergic synaptic plasticity in vitro, to impair long-term behavioral scores. In this scoring system, high anxiety index values aversive memory and long-lasting anxiety-related responses in vivo. represent an elevated anxiety-like behavior expression. The following formula was applied: 2. Material and methods anxiety index = 1- [(time spent in the open arms/ total time in the maze) + (number of entries to the open arms/ total time of maze exploration)]/2. 2.1. Graphene oxide material synthesis and characterization Exploratory and locomotor activities of rats (n = 6 per group) were s-GO solution was prepared by the modified Hummers‘ method as measured by the open field (OF) apparatus, a square arena with the 60 previously described [20], using graphite powder (Sigma-Aldrich) as a × 60 cm transparent plexiglass walls and the floor divided into 16 sec­ starting material [21] and under endotoxin-free conditions [22]. The tions (15 × 15 cm). Number of crossings (considered when the rat physicochemical characterization of the s-GO sheets was performed as crossed the border line of each section with four paws) were analyzed previously described [23] and is summarized in Table S1. following the EPM testing. All behavioral tests were performed under 40 Scanning electron microscopy (SEM) was performed using a FEI lx luminosity and videorecorded for off-line analysis. The XPloRat Magellan 400 L XHR SE microscope equipped with a newly developed software [25] was used to score the behaviors. Animals were anaes­ electron column with a monochromator, UC (UniColore) Technology. thetized with intraperitoneal injection of ketamine (Ketamine Imal­ Landing energy of 20.00 kV and beam current of 0.10 nA were used and gene®, Merial Laboratories) and xylazine (Sedaxylan®, Dechra signal was acquired with secondary electrons through-lens detector Veterinary Products) at 92 mg/kg and 10 mg/kg body weight, respec­ (TLD). 20 μL of s-GO solution at 100 μg/mL were deposited on an ul­ tively, and fixed in a stereotaxic frame. A stainless-steel guide cannula trathin carbon film on lacey carbon TEM grid and dried for at least 24 h. (outer diameter, 0.6 mm, and inner diameter, 0.4 mm) was implanted in Lateral dimension distribution was performed by manual counting of the the diencephalon aimed to the LA. The upper incisor bar was set at 3.3 longest sheet dimension on Image J software. The oxidation of the mm below the interaural line so that the skull was horizontal between material has been studied by several techniques such as Raman spec­ bregma and lambda. The guide cannula was vertically introduced using troscopy, surface charge (ζ-potential) measurements, TGA and XPS bregma as the reference and the following coordinates: A.P.− 3.48 mm, analysis (Table S1). M.L.− 5.2 mm and D.V.− 7 mm, according to Ref. [26]. At the end of the Animals, overall experimental design and surgery: All experimental surgery, the acrylic resin and two stainless steel screws were used to fix procedures were carried out in accordance with the Italian law (decree the guide cannula in the skull. In order to protect the guide cannula from 26/14) and the EU guidelines (2007/526/CE and 2010/63/UE) and obstruction a stainless-steel wire was used to seal it. Three days later, were approved by the Italian Ministry of Health (n. 689/2017-PR). Male rats were gently wrapped in a cloth and held while they received a adult Wistar rats weighed 230–250 g (n = 60) were used to perform the random treatment into LA of either s-GO (50 μg/mL) or ACSF solution in vivo experiments. Food and water were provided at libitum. The (composition described below) delivered by a needle (0.3 mm of outer enclosure was maintained at 20 ± 2 ◦ C on a light-dark cycle (lights on diameter) linked to a syringe (Hamilton) through a polyethylene tube. from 7 a.m. to 7 p.m.). Behavioral experiments were performed between The injection needle was inserted through the guide-cannula until it 9 a.m. and 2 p.m. Rats undergoing surgical procedures were deeply reached the LA (1 mm below the guide-cannula). 48 h later animals were anaesthetized. Analgesic and antibiotic medications were administrated submitted to the behavioral testing. postoperatively. All experimental procedures were planned to minimize the number of animals used and their suffering. We evaluated the aversive memory by the odor avoidance box, which consisted of a 2.2. Histology, microscopy and image analysis rectangular arena (40 × 26 × 36 cm) with black acrylic-plexiglass walls covered with a transparent plexiglass lid. At one side of the arena, an Rats were anaesthetized as described above and submitted to the alligator clip fixed in the wall is positioned 4 cm above the floor. In the transcardial perfusion with 0.1 M PBS followed by 4% formaldehyde opposite direction, a smaller box (20 × 26 × 22 cm) covered with a black (prepared from fresh paraformaldehyde; PFA, Sigma, St Louis, MO, plexiglass lid is positioned, named hide box. The arena and the hide box USA) in PBS. The brain was removed from the skull and prepared as were separated by a small 6 × 6 cm square hole allowing free access to described previously [27]. Briefly, brains were post fixed, cryoprotected, both chambers. Rats were placed (10 min) inside the hide box with free frozen on dry-ice and sectioned on a cryostat (Microm HM 550, Thermo access to the arena for 3 consecutive days to habituate to the apparatus. Fisher Scientific). The brain slices (30 μm coronal sections) were On the fourth day, the time spent in the following defensive behavior mounted on glass slides and processed according to either cresyl violet was recorded: head out (namely, the rat scanning the environment from staining or immunohistochemical methods. Brain slices were stained in a protected position, measured as poking of the head, or of head and 0.1% cresyl violet (Sigma) solution for 5 min, rinsed in distilled water shoulders, outside of the hide box but with the bulk of the rat body inside twice for 2 min each wash and dehydrated in a graded ethyl alcohol of it). Rats were divided in two groups (n = 6 per group), exposed to a series: 50%, 70%, 95% and 100%. Thereafter, they were cleared in piece (2 cm) of an unworn collar (UC), without any cat odor, and the xylene for 3 min and mounted with Permount (Fisher Scientific). His­ group exposed to the collar previously worn by the cat, named worn tological sections were analyzed and images were acquired using a Leica collar (WC). Collars were worn by an encaged cat. Rats were re-exposed DM6000 upright microscope with a 2.5 × dry objective. Alternatively, (10 min) to the context, arena without the cat collar to evaluate the brain slices were washed out with 0.1 M PBS, incubated with a blocking aversive memory related to the conditioned fear. Behaviors were solution composed of 3% BSA, 3% FBS, and 0.3% Triton X-100 in PBS for analyzed during the re-exposure to the context at 2 days and 6 days post- 45 min at room temperature (RT). The primary antibodies anti-ionized exposure. Long-term anxiety-related behavior was measured using the calcium-binding adaptor molecule 1 (Iba1; Wako, specific marker for 2 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 microglia, 1:500) and mouse anti-glial fibrillary acidic protein (GFAP; coordinates: Bregma − 1.8 mm, − 2.4 mm and 2.8 mm [34]. From these, Sigma-Aldrich, specific marker for glial cells, 1:400, diluted in saline and the amygdaloid complex was visually identified under a dissection mi­ 5% fetal bovine serum), were incubated overnight. Sections were croscope and dissected using a biopsy punch. The collected tissue was washed 3 times for 5 min each, incubated with the secondary antibodies enzymatically and mechanically dissociated and cells were seeded onto (1:400) AlexaFluor 488 goat anti-mouse (Thermo Fisher Scientific) and poly-L-ornithine-coated glass coverslips at a density of 800 cells/mm2 AlexaFluor 594 goat anti-rabbit, (Thermo Fisher Scientific), for 3 h in and maintained in controlled conditions (at 37 ◦ C, 5% CO2) for 8–12 the dark and washed. Nuclei were labeled with DAPI (Invitron, 1:500) in days prior to voltage-clamp recordings. PBS for 25 min at RT. All slides were cover slipped Fluoromounting [18]. Using a Nikon A1R confocal microscope, equipped with argon/ 2.5. s-GO treatments in vitro krypton, helium/neon, and UV lasers images were acquired with a 40 × (1.2 NA) objective and confocal sections were acquired every 0.5 μm up Acute slices were incubated for 5 h with 50 μg/mL of s-GO, added to a total Z-stack thickness of 20 μm. Gain, pixel-resolution, and expo­ directly to the ACSF, and as control, sister slices were kept in ACSF alone sure time remained constant for all images. From each section, we (saline-treated). At the end of incubation time, an individual slice was selected six regions of interest (ROI, 320 × 320 μm2), where the cell transferred to a submerged recording chamber and continuously densities of Iba1 and GFAP positive cells were calculated and normal­ perfused at 33–34 ◦ C with ACSF. In the experiments in which s-GO were ized for those of the contralateral hemisphere. In order to investigate the applied acutely, a gravity driven perfusion system was used and a so­ presence of s-GO in the LA 48 h after injection, the same brain slices lution of 10 μg/mL s-GO was applied at a speed of 2 mL/min for 5 min. In immunolabelled for DAPI, GFAP and Iba1 were visualized using the the experiments on dissociated cultures, s-GO were applied at a con­ reflection mode property during the confocal acquisition [28] at 40 × centration of 20 μg/mL for 30 s through the perfusion system in com­ (1.2 NA) with ROI of 140.77 × 140.77 μm2. bination with 50 μM of glutamate. Image analysis was performed using Volocity software (Volocity 3D image analysis software, PerkinElmer, U.S.A.). 2.6. Electrophysiology 2.3. Immunohistochemistry and golgi-staining and quantification of Patch clamp whole-cell recordings were obtained with glass pipettes dendritic spines (resistance 5–7 MΩ) filled with the following solution (in mM) 120 K gluconate, 20 KCl, 10 HEPES, 10 EGTA, 2 MgCl2, 2 Na2ATP (pH 7.3, After behavioral experiments, a group of rats (n = 8) re-exposed to osmolarity adjusted to 300 mOsm). The extracellular solution was ACSF the context (6 days) were anaesthetized as described above and decap­ for acute slices, and for dissociated cultures, the following saline solu­ itated, their brains were collected and prepared to the Golgi-Cox stain­ tion (in mM): 150 NaCl, 4 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, ing protocol [29]. Briefly, brains were incubated in the Golgi-Cox pH 7.4. Principal neurons of the lateral amygdala were visually identi­ solution (1% potassium dichromate, 0.8% potassium chromate and 1% fied thanks to their typical pyramidal shape [35] with an upright mi­ mercuric chloride) in the dark at RT for 25 days. After that, brains were croscope (Eclipse FN1; Nikon, Japan) equipped with differential incubated in sucrose solution at 30% for 24 h and sectioned in coronal interference contrast optics and digital videocamera (Nikon, Japan). sections (400 μm thickness) at the level of the amygdala using a vibra­ Data were collected by Multiclamp 700 A patch amplifier (Axon CNS, tome (Leica VT100S). Brain slices were mounted onto microscope slides Molecular Devices) and digitized at 10 KHz with the pClamp 10.6 with Permount (Fisher Scientific). Histological sections were analyzed acquisition-software (Molecular Devices LLC, USA). Membrane poten­ and images stacks of the LA neurons were acquired using a Leica tial values were not corrected for the liquid junction potential that was DM6000 upright microscope with a 63 × oil immersion objective. The of ~15 mV (calculated with the Clampex software; Molecular Devices, serial section images were aligned and dendritic spines of pyramidal Sunnyvale, CA, USA). The stability of the patch was checked by repet­ neurons were blind analyzed [30] in RECONSTRUCT software (http: itively monitoring the input and series resistance during the experi­ //synapses.clm.utexas.edu/tools/reconstruct/reconstruct.stm; RID: ments. Cells exhibiting 15% changes were excluded from the analysis. SCR_002716). Dendrite originating directly from the cell body with a The series resistance was <20 MΩ and it was not compensated. Input length of 70 μm from its origin was analyzed [31,32]. All dendrite resistance and cells capacitance were measured online with the mem­ protrusions were considered as spines, regardless to their morphological brane test feature of the pClamp software. In pyramidal neurons in the characteristics. LA slices, cell capacitance, input resistance and resting membrane po­ tential did not show any statistically significant difference between in 2.4. In vitro amygdala preparations s-GO treated and saline treated neurons (p > 0.05, Table S2). In current clamp experiments, 1 s long lasting steps of positive cur­ All experimental procedures were carried out in accordance with the rent with increasing amplitude (10, 20, 30, 40, 50, 60, 70, 80, 90, 100 Italian law (decree 26/14) and the EU guidelines (2007/526/CE and pA) were injected in pyramidal cells from a resting membrane potential 2010/63/UE) and were approved by the Italian Ministry of Health (n. of − 65 mV. Both s-GO-treated and untreated neurons showed compa­ 689/2017-PR, n. 22DAB.N.1Z8 and n. 22DAB.N.1WO). For acute slices rable firing properties (Fig. 4C), measured as threshold for action po­ male P12–P16 juvenile Wistar rats (n = 17) were used. Briefly, after tential, amount of injected current needed to reach it and action being decapitated, the brain was quickly removed from the skull and potential amplitudes. No statistically significant differences were placed in ice-cold solution containing (in mM): sucrose 215, KCl 3.5, observed between the two groups (p > 0.05, Table S2). NaH2PO4 1.2, NaHCO3 25, CaCl2 2, MgCl2 1.3, glucose 25 and ascorbic In the experiments where slices were chronically incubated in s-GO, acid 4, saturated with 95% O2–5% CO2 (pH 7.3–7.4). Coronal amygdalar the glutamatergic synaptic activity was recorded at a holding potential slices (300 μm thick) were cut with a vibratome (LeicaVT1000S) and of - 65 mV, in the presence of gabazine (10 μM; a blocker of GABAA stored at room temperature in a holding bath containing 130 mM NaCl receptor mediated postsynaptic currents). EPSCs were collected from instead of sucrose that was gassed with 95% O2–5% CO2 (artificial ce­ traces and averaged (showed in Fig. 4F) and no differences were rebrospinal fluid solution, ACSF). observed in their amplitude (in saline-treated neurons: 9 ± 1 pA and in s- For cultures, postnatal P7-10 Wistar rats were used. Dissociated GO-treated ones: 8 ± 2 pA), rise time (in saline-treated neurons: 1.2 ± amygdalar cultures were prepared as previously described [33]. Briefly, 0.2 ms and in s-GO-treated ones: 1.1 ± 0.1 ms) and decay time constant τ brains were quickly removed and cut in coronal sections (400 μm (in saline-treated neurons: 5.4 ± 0.9 ms and in s-GO-treated ones: 4.7 ± thickness) using a vibratome (Leica VT100S). Three sections at the level 0.4 ms) when incubated with s-GO (n = 8 cells) in respect to control (n = of the amygdalar complex were collected using the following 9 cells; p > 0.05). After recording spontaneous activity for 8 min as 3 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 baseline, LTP was induced in dissociated cultures by applying 50 μM of were checked for normality and homogeneity and analyzed using Stu­ glutamate for 30 s, while the membrane potential of the recorded cell dent’s unpaired two-tailed t-test. All comparisons between more than 2 was depolarized from − 58 mV to +4 mV (under voltage clamp mode). or 3 groups were made with one-way ANOVA and two-way ANOVAs, The effects of LTP induction were monitored for 30 min by measuring respectively, followed by Bonferroni post hoc test. For electrophysio­ EPSC and IPSC frequencies and amplitudes. The values reported are logical data, Shapiro-Wilk normality test was applied to evaluate the averages calculated between 24 and 30 min after LTP induction and statistical distribution of the data sets. Statistically significant difference normalized for the pre-treatment baseline values. In the experiments of between two data sets was assessed by one-way ANOVA (if distributed as s-GO application through the perfusion system in acute slices and in a Gaussian distribution) or by Mann–Whitney test (if not). In experi­ dissociated cultures, sPSCs were recorded in the absence of any drugs ments obtained from dissociated amygdalar cells, for parametric data, and were analyzed offline using the software AxoGraph X (Axograph the statistically significant difference among the three groups was Scientific), which exploits a detection algorithm based on a sliding assessed through one-way ANOVA, using Holm-Sidak’s multiple com­ templates to separate EPSCs and IPSCs on the basis of their different parisons test for post hoc analysis. Not parametric data were analyzed decay times (see Results). For each recording, events were collected and with Kruskal-Wallis test and post hoc analysis was done with Dunn’s averaged to measure the peak amplitude and kinetic properties on the multiple comparison test. All values are expressed as mean ± SEM. P- resulting trace. The decay time of PSCs was calculated by fitting the value < 0.05 was considered statistically significant and n is the number decaying phase of the current with a mono-exponential function. In of animals if not otherwise stated. paired recordings, the presynaptic neuron in current clamp mode was held at − 70 mV (by ≤ 0.02 nA negative current injection), and action 3. Results potentials were evoked by delivering short (4 ms) positive (1 nA) square current pulses. Monosynaptic connections were identified by their short The produced biology-grade s-GO dispersion [18] was characterized delay (<3 ms) [18], measured between the peak of the evoked action by AFM and SEM; in both measures s-GO flakes displayed a similar range potential and the onset of the unitary evoked PSCs. Recordings of EPSCs of lateral dimensions: 0.06–1.7 μm by AFM (Fig. 1A and B) and 0.10–2.0 and IPSCs at different holding potentials were used to extrapolate I/V μm by SEM (Fig. 1D and E). Fig. 1C shows s-GO thickness, that ranged curves (Supplementary Figure S2). between 1 and 1.5 nm in accordance with single or few-layer sheets. The oxidation of the material was analyzed by Raman spectroscopy, surface 2.7. Data analysis and statistics charge (ζ-potential) measurements, TGA and XPS analysis, these mea­ sures are summarized in Table S1. Data from independent groups of animals exposed to the cat collar Structural and morphological characterization of s-GO sheets by Fig. 1. s-GO AFM and SEM features and the behavioral model of short- and long-term fear memories induced by aversive olfactory stimulus. 4 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 AFM: (A) height image; (B) lateral dimension distribution (based on 123 odor avoidance box (sketched in Fig. 1F) to quantify the emergence of sheets); (C) cross section analysis highlighted by the dashed line in the defensive behavior (“head out” [40]; see methods). We compared height image; and by SEM: (D) micrograph; (E) lateral dimension dis­ defensive behavior in two groups of rats, one group exposed to a tribution (based on 99 sheets). (F) Sketched experimental settings and cat-worn collar (WC; n = 12) and a control group exposed to an unworn timeline. Upon habituation in the avoidance box, adult rats were collar (UC; n = 12). In Fig. 1G the bar plots show the significant (t (10) = exposed to either worn (WC) or unworn (UC) cat collar. 2 and 6 days 7.01, p < 0.001) increase in “head out” behavior when comparing WC after the exposure, rats were re-exposed to the context. (G) Bar plots with UC exposed animals. The avoidance fear response was also signif­ summarize the increase in the time of head out defensive response in rats icantly increased when comparing UC with WC groups re-exposed to the caused by either the cat odor exposure or the re-exposure to the context context after 2 days (t (10) = 7.46, p < 0.001) and after 6 days (t (10) = 2 or 6 days later, namely short- and long-term aversive memory, 3.13, p < 0.01), suggesting the induction of short- and long-term aver­ respectively. N = 6 each group. ***p < 0.001; **p < 0.01 versus UC sive memories, related to PTSD emergence [41,42]. The emergence of groups. PTSD has been associated to augmented dendritic spines in the amyg­ To test thin s-GO efficacy to affect neurotransmission in the LA, dala nuclei [31,32]. This anatomical signature of synaptic plasticity was preventing excitatory synapses reinforcement and anxiety disorders, we confirmed in the amygdala of WC groups (6 days after first exposure, n established a model of PTSD upon rat exposure to an aversive stimulus, = 4) where the number of dendritic spines in LA was significantly (p < the predator (cat) odor, known to activate the amygdaloid complex [36, 0.05) higher when compared to UC groups (n = 4; Supplementary 37]. Such a treatment leads to hyperexcitability of LA circuits, fear re­ Figure S1). sponses and altered behavior [38,39]. In adult rats, immediate innate fear response was induced by single experience of an aversive (predator) 3.1. s-GOs locally delivered in the LA impair long-term fear memories olfactory stimulus and the emergence of long-lasting specific anxiety behavior was evaluated by re-exposure to the context. We adopted an The contextual avoidance behavior to predatory odor is ultimately Fig. 2. Lateral amygdala (LA) treatment by s-GO local injection. (A) Schematic representation of experimental timeline showing the sequence of the experimental procedures, from habituation to stereotaxic surgery and histology. (B) Top, diagram of the rat brain showing the target area for s-GO or saline microinjections. Bottom, (i) representative photomicrograph of Nissl-staining (yellow arrow indicates the site of s-GO injection); scale bar 300 μm. In (ii, iii, iv, v, vi) schematic drawings of rat brain sections summarizing the microinjection points of either s-GO (* when in the LA, # when in the perirhinal cortex) or saline (@ when in the LA, + when in the perirhinal cortex). Each symbol represents an independent experiment made on a single rat. (C) Confocal photomicrographs showing GFAP- (green), Iba1- (red) reactivity and DAPI- (blue) stained nuclei in the LA slices of rats treated with either saline (left) or s-GO (right). (D) The bar plot summarizes the tissue responses in saline or s-GO treatments, N = 6 slices each group. *p > 0.05 according to the unpaired t-test. Scale bar 50 μm. (E) Representative images of LA slices (48 h after injection; same samples as in C, different fields) with DAPI (in blue) and s-GO sheets (in yellow) visualized by the reflection mode of the confocal system. Note that s-GO is visible as aggregates only in s-GO injected animals. Scale bar 20 μm. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) 5 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 organized by the LA glutamatergic neurons [43]. We tested the hy­ 2 days after delivery of s-GO or saline in WC and UC groups. During the pothesis that s-GO injection into the LA, via transient impairment of re-exposure to the stress-related context, in LA saline-treated animals, glutamatergic synapses [18], may prevent PTSD-related behavior we detected the expected (see also Fig. 1G) significant increase in the emergence in rodents. Such an effect would fortify the strategy of se­ “head out” response in WC (F (1,20) = 0.04, p < 0.001) when compared lective and transient targeting of synapses to prevent the development of to UC group. Remarkably, in LA s-GO-treated animals, flakes injection brain pathologies. reversed the long-term conditioned fear reaction in WC (F (1,20) = 19, p Two independent groups of rats (n = 36) were exposed to WC or UC < 0.01) when compared to LA saline-treated WC (p > 0.05; Fig. 3A and context and, after 24 h, a guided cannula was stereotaxically implanted B) and s-GO effect was lost when delivered outside the LA, in the close- into the rat brain, targeting the LA or, alternatively, the perirhinal cor­ by perirhinal cortex (P > 0.05; Fig. 3B). To evaluate the effective ability tex, to deliver locally, three days later, either s-GO (50 μg/mL) [18] or of s-GO to remove long-term aversive memory and thus PTSD-associated the vehicle (saline solution). Two days after s-GO or saline delivery, WC behavior in WC upon s-GO injections, we quantified rat behavior in UC, and UC groups of animals were re-exposed to the context (i.e. 6 days WC saline and WC s-GO, using the elevated plus maze (EPM) and open after the first exposure) and then euthanized (sketched in Fig. 2A). The field (OF) apparatus (Fig. 3B). position of the cannula was confirmed histologically in all animals both We quantified PTSD-like behavior by measuring the anxiety index in when targeting the LA and when targeting the perirhinal cortex, a the EPM. WC rats, injected with saline, exhibited a significantly higher neighboring brain structure (see Methods, schematic drawings and (F (91,20) = 0.88, p < 0.05) anxiety index 6 days after the exposure to histology of Fig. 2B). Tissue reactivity was studied by immunofluores­ the predator odor respect to that of UC (saline treated) group. On the cence assessment of GFAP-positive astrocytes and Iba1-positive micro­ contrary, s-GO microinjection into LA reduced the PTSD related glia. At 6 days post-surgery, tissue reactivity was comparable in the two behavior in WC animals, decreasing the anxiety index in a statistically animal groups, saline or s-GO treated (Fig. 2C and bar plot in D). significant manner (F (1,20) = 13.43, p < 0.01), if compared to that of We tested the presence of s-GO within the LA (48 h post-injection) by the WC saline treated rats. In addition, WC rodents treated with s-GO did operating the confocal microscopy under reflection mode, which allows not differ from UC (saline or sGO injected) animals (Fig. 3B). We ruled the visualization of s-GO [28]. Fig. 2E shows confocal reconstructions of out mere alterations in animal locomotion due to s-GO by testing loco­ saline and s-GO treated LA, in the latter residual aggregates of s-GO (in motor behavior via the OF apparatus (p > 0.05; Fig. 3B). yellow, reflection mode) were detected, confirming our previous ob­ servations on the duration of s-GO permanence once injected in the adult brain [18]. 3.2. s-GO flakes decrease the activity of glutamatergic synapses in the LA In all animals, aversive behavior prior to surgery was quantified and the avoidance response was significantly higher in the WC group when Suggested molecular mechanisms of the fear components that un­ compared to UC one (t (10) = 4.09, p < 0.05; Fig. 3A and B). We then derlie PTSD include synaptic long-term potentiation (LTP) in LA gluta­ analyzed long-term aversive memory related behavior at day 6, namely matergic synapses [44]. We hypothesize that s-GO, once injected in the LA, transiently impair local excitatory glutamatergic synapses, as Fig. 3. s-GO delivery in LA impairs long-term fear memories. (A) Schematic representation of the experimental timeline and behavioral testing (at day 6). (B) On the left, bar plot summarizing the head out behavioral responses evoked by the exposure to UC or WC and by the re-exposure to the context after s-GO or saline mi­ croinjections into either the lateral amygdala (LA) or the perirhinal cortex. In the middle, bar plot showing the anxiety index evaluated in the EPM test in UC and WC treated with s-GO or saline. On the right, bar plot reporting OF test in UC and WC treated with s-GO or saline. N = 6 for each group. ***p < 0.001; **p < 0.01; *p < 0.05. 6 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 Fig. 4. s-GO downregulate the activity of glutamatergic synapses in the LA (A) Sketch of the experimental settings. (B) Current clamp traces of elicited firing activity in control neurons and in s-GO treated ones. (C) Plot reporting the number of APs fired versus the current injected, for saline-treated (n = 9) and s-GO-treated cells (n = 8). (D) On the left, voltage clamp traces of spontaneous EPSCs recorded in the presence of gabazine (10 μM) in control neurons and s-GO treated ones. On the right, averaged EPSCs waveforms (same cells as left). (E) Cumulative probability plot of the inter event interval of EPSCs for control and s-GO treated neurons: s-GO shift the curve toward larger values. (F) Voltage clamp recordings of sPSCs during acute application of s-GO (n = 9 cells) or saline as control (n = 8 cells) through the perfusion system. Note the downregulation of neuronal activity in s-GO treated cell, while no effect was observed in control. (G) Bar plot showing the statistically significant decrease in normalized sPSCs frequency in s-GO treated cells respect to controls. (H) Averaged EPSCs and IPSCs (same cells as in G) isolated offline on the basis of their different decay times. (I) Bar plots of normalized EPSCs and IPSCs frequency showing that s-GO affect specifically excitatory synapses. Normalized EPSC frequency was decreased (p = 0.0016) during the treatment and during the early phase of the wash out (p = 0.036). Diversely, the normalized IPSCs frequency was not changed by s-GO application. observed in previous studies [18,19]. Thus, mechanistically, s-GO currents (sPSCs) in the absence of pharmacological blockers. Fig. 4F–G reduction in glutamatergic synaptic efficacy might prevent LTP shows that s-GO significantly reduced sPSCs frequency when compared expression, crucial to PTSD development. We challenged this hypothesis to saline (for saline: 0.87 ± 0.06 and for s-GO: 0.47 ± 0.09 Hz; p = by single cell electrophysiology in vitro. We first incubated acute brain 0.0053) and this reduction was maintained during the early phase of explants containing the amygdala complex with saline or s-GO (50 s-GO removal (for saline: 0.89 ± 0.06 Hz; and for s-GO: 0.51 ± 0.14; p = μg/mL, in saline, 5 h; sketched in Fig. 4A). Under current clamp mode, a 0.0297), with a partial recovery after 10 min washout (for saline: 0.95 ± homogeneous population of neurons was recorded, identified by their 0.16 and for s-GO: 0.55 ± 0.16; p > 0.05). Fig. 4H–I shows that in the passive (see methods and Table S2) and active (Table S2, Fig. 4B and C) same experimental conditions, s-GO specifically targeted EPSCs and was membrane properties, which were not altered by s-GO treatments. ineffective on inhibitory GABAA receptor mediated postsynaptic cur­ In voltage clamp mode the inter-event interval of spontaneous rents (IPSCs), identified by their kinetic properties (shown in Fig. 4H, excitatory glutamatergic postsynaptic currents (EPSCs), pharmacologi­ EPSCs displayed typical fast ~5 ms day, while IPSCs featured a slower cally isolated in the presence of gabazine (10 μM; see methods and ~13 ms one [18]) and pharmacology (10 μM CNQX readily abolished Fig. 4D) were increased after s-GO treatments (432 ± 10 ms) in respect fast event, while 10 μM gabazine removed slow ones). Such analysis to controls (326 ± 5 ms; see cumulative plot in Fig. 4E; P value < confirmed that the effect of s-GO in LA tissue was specific for gluta­ 0.0001). This finding suggests that LA ex-vivo slices, upon prolonged matergic synapses [18,19]. exposure to s-GO [18], displayed a downregulation of glutamatergic transmission. To gain more insights on s-GO dynamics, we administered the nanomaterial acutely (10 μg/mL, 5 min) to the amygdala brain ex­ plants while monitoring heterogeneous spontaneous post synaptic 7 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 3.3. In vitro long-term potentiation of amygdala synapses is impaired by s- 4. Discussion GO We describe here the efficacy of s-GO in preventing the reinforce­ The rescue of anxiety related behaviors observed in animals injected ment of aversive memory and the development of long-term anxiety with s-GO hinted at the nanomaterial interference with the building up behavior only when targeting synapses in the LA. We propose that this of the pathological LTP induced by the odor exposure [45]. To explore effect is due to the ability of s-GO to transiently inhibit glutamatergic this possibility, we adopted a simplified in vitro model of LTP induction activity in LA excitatory circuits in a synapse-specific and localized in the amygdala, which allows direct experimental control and access to manner, ultimately preventing the emergence of LA dysfunctional glutamatergic synapses. We induced synaptic plasticity in dissociated plasticity. amygdala cultures (Fig. 5A–C) by coupling a brief (30 s) application of We used a widely accepted [36,37] animal model of PTSD. This glutamate (50 μM) with a depolarization (+4 mV) of the neuronal model is characterized by a correlation between hyperactivity in the LA membrane, favoring the activation of glutamate NMDA receptors, glutamatergic circuitry and the emergence of anxiety disorders [39,48]. involved in LTP induction [46,47]. Such a treatment resulted in a 30 min In our experiments, the aversive stimulus induced an innate fear long lasting and stable increase in the amplitude of EPSCs (identified by response, measured as an increase in the head out behavior [49]. The rat their kinetic properties, voltage dependence and pharmacology, as maintenance of such a behavior when re-exposed to the cat shown in the Supplementary Figure 2) in respect to the baseline values, odor-associated context (i.e. neutral environment) [49] supported the while in the controls, that underwent to membrane depolarization in the fear memory consolidation and, in animals tested six days after the first absence of glutamate application, no changes were detected (normal­ exposure to the stressful event, the development of long lasting anxiety, ized amplitudes were 0.81 ± 0.09 in controls, n = 14; 1.72 ± 0.21 in as indicated by the EPM test [42]. Long-term aversive memory and the glutamate treated cells, n = 20; p = 0.0004, Fig. 5B–D, 30 min). s-GO development of long-term anxiety, evoked by a predator odor stress, (20 μg/mL, 30 s) applied simultaneously to glutamate, blocked the EPSC represent a model of PTSD [50]. In PTSD patients an involvement of potentiation (p = 0.0007, Fig. 5C–D). EPCSs frequency remained unal­ amygdala subregion dysfunctional connectivity has been reported [51], tered in all treatments (Fig. 5E). IPSCs amplitude (Fig. 5 D) and fre­ accordingly, in the experimental PTSD rat model, neuronal circuitry in quency (Fig. 5E) were not modulated by these treatments. the LA undergoes to a reshaping of excitatory glutamatergic synapses These findings indicate that s-GO target specifically glutamatergic [52] featuring the progressive building up of LA synaptic potentiation synapses in LA and further suggest that, when applied in conjunction [53]. Such a boosted excitatory neuronal activity is proposed to trigger with the LTP inducing stimuli, s-GO may prevent the development of the long-term anxiety related behavior [54]. glutamatergic synaptic plasticity. s-GO has been reported to specifically interact with glutamatergic Fig. 5. s-GO impair the long-term potentiation of amygdala synapses in vitro. (A) Schematic representation of amygdala dissociated cultures and bright field image of cultured cells. (B) Representative traces of neuronal spontaneous synaptic activity showing LTP induction in amygdala cultures and its impairment upon s-GO treatment. Chemical LTP was obtained by coupling the depolarization of postsynaptic membrane potential with the application of glutamate at the concentration of 50 μM for 30 s (n = 20 cells). In s-GO treated cells, the nanomaterial was applied together with glutamate under the same condition (n = 17 cells), while control cells were perfused with saline (n = 14 cells). Note the lack of potentiation in s-GO treated cell. (C) Offline analysis of sPSCs isolated EPSCs and IPSCs, before and after the treatments. Note that only upon glutamate treatment there is a potentiation of neuronal activity detectable as an increment of the EPSCs amplitude. (D) Plots showing that s-GO treatment blocks the 30 min long lasting increase in EPSCs amplitude, observed in neurons undergone to the LTP induction protocol. (E) Bar plots showing that the frequency of EPSCs or IPSCs was not affected either by LTP induction or s-GO application. *p < 0.05. 8 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 synapses in the CNS inducing a downregulation of excitatory neuro­ the work reported in this paper. transmission when applied chronically in vitro at a concentration of 10 μg/mL [55] or delivered in vivo in one single application at a higher Acknowledgements dose. [18,19] In vivo s-GO inhibition of glutamatergic synapses in the spinal cord modulates motor behavior [19] and in the rat hippocampus This work has received funding from the European Union Horizon local delivery of s-GO significantly and selectively sized down gluta­ 2020 Research and Innovation Programme under Grant Agreement matergic activity for 48 h, without altering neuronal viability [18]. In numbers GrapheneCore2 (785219) and GrapheneCore3 (881603) and that study, we explored the time course of s-GO fate, which was found to from the São Paulo Research Foundation (FAPESP) number 2016/ match the synaptic silencing reversibility [18]. Based on these previous 18218–0. We would like to acknowledge the staff of the Electron Mi­ evidences, we injected a single dose of s-GO in the LA four days after the croscopy Unit at ICN2 for their expertise and assistance on the SEM. The exposure to the cat odor to impair glutamatergic transmission for 48 h ICN2 is funded by the CERCA programme, Generalitat de Catalunya, and and eventually interrupt the building up of glutamatergic synaptic is supported by the Severo Ochoa Centres of Excellence programme by plasticity. Indeed, the presence of s-GO in the treated amygdala at 48 h the Spanish Research Agency (AEI, grant no. SEV-2017-0706). post injection was confirmed and s-GO treatment was effective in dis­ rupting both long term aversive memory and anxiety related behaviors. Appendix A. Supplementary data We ruled out potential s-GO nonspecific effects, in fact anxiety responses were unperturbed after 48 h of s-GO delivery outside the LA. In addition, Supplementary data to this article can be found online at https://doi. we never detected motor alterations in s-GO treated animals, potentially org/10.1016/j.biomaterials.2021.120749. affecting nonspecifically rat defensive behavior. In this framework, also a generic impact of the surgery procedure or an increased tissue reac­ Data and materials availability tivity brought about by s-GO injections and affecting synaptic rein­ forcement, were excluded. The data that support the findings of this study are available from the We adopted in vitro amygdala models to experimentally test the corresponding authors upon reasonable request. mechanistic interaction between s-GO and excitatory synapses. In acute amygdala slices we confirmed that chronic (mimicking in vivo accu­ References mulation) or sub-acute (mimicking in vivo diffusion) exposures to s-GO both resulted in a specific reduction in EPSCs frequency, leaving IPSCs [1] E.R. Duval, J. Sheynin, A.P. King, K.L. Phan, N.M. Simon, B. Martis, K.E. Porter, S. unchanged, in accordance to previous reports where s-GO was tested in B. Norman, I. Liberzon, S.A.M. Rauch, Depress. Anxiety 37 (2020) 670–681. different CNS areas [18,19,55]. s-GO has been suggested to target pre­ [2] A.L. Gold, R. Abend, J.C. Britton, B. Behrens, M. Farber, E. Ronkin, G. Chen, E. Leibenluft, D.S. Pine, Am. J. Psychiatr. 177 (2020) 454–463. synaptic glutamate release in the hippocampus [18], a feature in prin­ [3] J.E. Hassell Jr., K.T. Nguyen, C.A. Gates, C.A. Lowry, Curr Top Behav Neurosci 43 ciple enabling the alteration of presynaptic mechanisms and (2019) 271–321. glutamatergic transmission engaged in LTP. We further tested in [4] N.C. Coimbra, T. Paschoalin-Maurin, G.S. Bassi, A. Kanashiro, A.F. Biagioni, T. T. Felippotti, D.H. Elias-Filho, J. Mendes-Gomes, J.P. Cysne-Coimbra, R.C. Almada, cultured amygdala circuits [33] whether s-GO could impair gluta­ B. Lobão-Soares, Rev. Bras. Psiquiatr. 39 (2017) 72–83. matergic synaptic potentiation. We successfully potentiated EPSCs by [5] P. Tovote, J.P. Fadok, A. Lüthi, Nat. Rev. Neurosci. 16 (2015) 317–331. chemical LTP paradigm (cLTP) [46,56–58] and the application of s-GO [6] D. Busti, R. Geracitano, N. Whittle, Y. Dalezios, M. Mańko, W. Kaufmann, K. Sätzler, N. Singewald, M. Capogna, F. Ferraguti, J. Neurosci. 31 (2011) during cLTP induction counteracted the synaptic potentiation. Although 5131–5144. we cannot demonstrate that in vivo s-GO prevented the potentiation of [7] R.K. Butler, A.C. Sharko, E.M. Oliver, P. Brito-Vargas, K.F. Kaigler, J.R. Fadel, M. amygdala circuits usually linked to PTSD development [44], it is A. Wilson, Bone 23 (2011) 133–144. [8] S. Duvarci, D. Pare, Neuron 82 (2014) 966–980. tempting to speculate that the injection of s-GO in the LA during a [9] M.P. Parsons, L.A. Raymond, Neuron 82 (2014) 279–293. temporal window crucial to the reinforcement of plastic changes related [10] D.M. Bannerman, R. Sprengel, D.J. Sanderson, S.B. Mchugh, J.N.P. Rawlins, to the contextual fear memory acquisition [38], effectively reduced H. Monyer, P.H. Seeburg, Nat. Rev. Neurosci. 15 (2014) 181–192. [11] L. Musazzi, G. Racagni, M. Popoli, Neurochem. Int. 59 (2011) 138–149. glutamatergic transmission potentiation and the onset of downstream [12] J.-Y. Zhang, T.-H. Liu, Y. He, H.-Q. Pan, W.-H. Zhang, X.-P. Yin, X.-L. Tian, B.-M. Li, long lasting anxiety-related behavior. s-GO nanoflakes might therefore X.-D. Wang, A. Holmes, T.-F. Yuan, B.-X. Pan, Biol. Psychiatr. 85 (2019) 189–201. be clinically exploited in the broader area of engineered nanoparticles [13] A.K. Geim, K.S. Novoselov, Nat. Mater. 6 (2007) 183–191. for precision-medicine applications [59], to allow improved sub-cellular [14] L. Feng, L. Wu, X. Qu, Adv. Mater. 25 (2013) 168–186. [15] G. Yang, J. Su, J. Gao, X. Hu, C. Geng, Q. Fu, J. Supercrit. Fluids 73 (2013) 1–9. (i.e. synaptic) targeting in neurological disorders. [16] K. Kostarelos, M. Vincent, C. Hebert, J.A. Garrido, Adv. Mater. 29 (2017) 1–7. [17] M. Bramini, G. Alberini, E. Colombo, M. Chiacchiaretta, M.L. DiFrancesco, J. 5. Conclusion F. Maya-Vetencourt, L. Maragliano, F. Benfenati, F. Cesca, Front. Syst. Neurosci. 12 (2018) 1–22. [18] R. Rauti, M. Medelin, L. Newman, S. Vranic, G. Reina, A. Bianco, M. Prato, Our experiments suggest that s-GO nanoflakes, thanks to their action K. Kostarelos, L. Ballerini, Nano Lett. 19 (2019) 2858–2870. as synaptic and behavioral modulators, might effectively hinder path­ [19] G. Cellot, S. Vranic, Y. Shin, R. Worsley, A.F. Rodrigues, C. Bussy, C. Casiraghi, K. Kostarelos, J.R. McDearmid, Nanoscale Horizons 5 (2020) 1250–1263. ological behaviors based on aberrant glutamatergic transmission. [20] H. Ali-Boucetta, D. Bitounis, R. Raveendran-Nair, A. Servant, J. Van den Bossche, K. Kostarelos, Adv. Healthc. Mater. 2 (2013) 433–441. Author contributions [21] D.A. Jasim, N. Lozano, K. Kostarelos, 2D Mater. 3 (2016), 014006. [22] S.P. Mukherjee, N. Lozano, M. Kucki, A.E. Del Rio-Castillo, L. Newman, E. Vazquez, K. Kostarelos, P. Wick, B. Fadeel, PloS One 11 (2016), e0166816. AFB and NCC designed the PTSD paradigm; AFB performed the [23] A.F. Rodrigues, L. Newman, D.A. Jasim, I.A. Vacchi, C. Ménard-Moyon, L.E. Crica, behavior, surgery and histology; RC performed confocal microscopy; GC A. Bianco, K. Kostarelos, C. Bussy, Arch. Toxicol. 92 (2018) 3359–3379. [24] H. Cohen, J. Zohar, Z. Kaplan, J. Arnt, Eur. Neuropsychopharmacol 28 (2018) designed the electrophysiology; GC and EP performed the electrophys­ 63–74. iology; AFB, GC and EP analyzed data; NL, BB and KK design, produced [25] J. Tejada, K.T. Chaim, S. Morato, Psicol. Teor. Pesqui. 33 (2017) e3322. and characterized materials; LB conceived the idea, designed research [26] G. Paxinos, C.R. Watson, The Rat Brain in Stereotaxic Coordinates, sixth ed., and wrote the MS. Elsevier Academic Press, San Diego, 2007. [27] A.F. Biagioni, R.C. de Oliveira, R. de Oliveira, J.A. da Silva, T. dos Anjos-Garcia, C. M. Roncon, A.P. Corrado, H. Zangrossi, N.C. Coimbra, Eur. Neuropsychopharmacol Declaration of competing interest 26 (2016) 532–545. [28] M. Musto, R. Rauti, A.F. Rodrigues, E. Bonechi, C. Ballerini, K. Kostarelos, L. Ballerini, Front. Syst. Neurosci. 13 (2019) 1–15. The authors declare that they have no known competing financial [29] E.L. Louth, C.D. Sutton, A.L. Mendell, N.J. Maclusky, C.D.C. Bailey, JoVE 122 interests or personal relationships that could have appeared to influence (2017), e55358. 9 A. Franceschi Biagioni et al. Biomaterials 271 (2021) 120749 [30] W.C. Risher, T. Ustunkaya, J.S. Alvarado, C. Eroglu, PloS One 9 (2014), e107591. [46] A. Malgaroli, R.W. Tsien, Nature 357 (1992) 134–139. [31] P. Chakraborty, S. Chattarji, Psychopharmacology 236 (2019) 73–86. [47] R.J. Cormier, P.T. Kelly, J. Neurophysiol. 75 (1996) 1909–1918. [32] R. Mitra, R. Adamec, R. Sapolsky, Behav. Pharmacol. 205 (2009) 535–543. [48] H.T. Blair, G.E. Schafe, E.P. Bauer, S.M. Rodrigues, J.E. LeDoux, Learn. Mem. 8 [33] N. Secomandi, A. Franceschi Biagioni, K. Kostarelos, G. Cellot, L. Ballerini, (2001) 229–242. Nanomed. Nanotechnol. Biol. Med. 26 (2020) 102174. [49] I.S. McGregor, L. Schrama, P. Ambermoon, R.A. Dielenberg, Behav. Brain Res. 129 [34] R. Khazipov, D. Zaynutdinova, E. Ogievetsky, G. Valeeva, O. Mitrukhina, J. (2002) 1–16. B. Manent, A. Represa, Front. Neuroanat. 9 (2015) 1–5. [50] M.A. Matar, H. Cohen, Z. Kaplan, J. Zohar, Neuropsychopharmacology 31 (2006) [35] J.P. Johansen, H. Hamanaka, M.H. Monfils, R. Behnia, K. Deisseroth, H.T. Blair, J. 2610–2618. E. LeDoux, Proc. Natl. Acad. Sci. U. S. A 107 (2010) 12692–12697. [51] D. Rabellino, M. Densmore, P.A. Frewen, J. Théberge, M.C. McKinnon, R.A. Lanius, [36] R.A. Dielenberg, G.E. Hunt, I.S. McGregor, Neuroscience 104 (2001) 1085–1097. PloS One 11 (2016), e0163097. [37] L.K. Takahashi, B.R. Nakashima, H. Hong, K. Watanabe, Neurosci. Biobehav. Rev. [52] B.M. Sharp, Transl. Psychiatry 7 (2017) e1194. 29 (2005) 1157–1167. [53] J.P. Johansen, C.K. Cain, L.E. Ostroff, J.E. LeDoux, Cell 147 (2011) 509–524. [38] N.S. Canteras, E. Pavesi, A.P. Carobrez, Front. Neurosci. 9 (2015) 1–10. [54] O. Babaev, C. Piletti Chatain, D. Krueger-Burg, Exp. Mol. Med. 50 (2018) 1–16. [39] J.A. Rosenkranz, E.R. Venheim, M. Padival, Biol. 67 (2010) 1128–1136. [55] R. Rauti, N. Lozano, V. León, D. Scaini, M. Musto, I. Rago, F.P. Ulloa Severino, [40] L.G. Staples, I.S. McGregor, R. Apfelbach, G.E. Hunt, Neuroscience 151 (2008) A. Fabbro, L. Casalis, E. Vázquez, K. Kostarelos, M. Prato, L. Ballerini, ACS Nano 10 937–947. (2016) 4459–4471. [41] A. Berardi, V. Trezza, M. Palmery, L. Trabace, V. Cuomo, P. Campolongo, Front. [56] R.J. Cormier, M.D. Mauk, P.T. Kelly, Neuron 10 (1993) 907–919. Behav. Neurosci. 8 (2014) 1–12. [57] A. Louveau, J. Angibaud, F. Haspot, M.C. Opazo, R. Thinard, V. Thepenier, S. [42] C. Muñoz-Abellán, R. Andero, R. Nadal, A. Armario, Psychoneuroendocrinology 33 J. Baudouin, L. Lescaudron, P. Hulin, C.A. Riedel, H. Boudin, J. Neurosci. 33 (2008) 1139–1150. (2013) 18672–18685. [43] R.C. Martinez, E.F. Carvalho-Netto, É.R. Ribeiro-Barbosa, M.V.C. Baldo, N. [58] M. Kintscher, C. Wozny, F.W. Johenning, D. Schmitz, J. Breustedt, Nat. Commun. 4 S. Canteras, Neuroscience 172 (2011) 314–328. (2013) 1–8. [44] A.L. Mahan, K.J. Ressler, Trends Neurosci. 35 (2012) 24–35. [59] M.J. Mitchell, M.M. Billingsley, R.M. Haley, M.E. Wechsler, N.A. Peppas, R. Langer, [45] K.J. Wallace, J.B. Rosen, J. Neurosci. 21 (2001) 3619–3627. Nat. Rev. Drug Discov. 20 (2020) 101–124. 10