Increased P2×2 receptors induced by amyloid-β peptide participates in the neurotoxicity in alzheimer’s disease

Biomedicine & Pharmacotherapy 142 (2021) 111968 Available online 31 July 2021 0753-3322/© 2021 The Authors. Published by Elsevier Masson SAS. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Increased P2×2 receptors induced by amyloid-β peptide participates in the neurotoxicity in alzheimer’s disease Pamela A. Godoy a , 1 , Daniela Mennickent a , 1 , Inmaculada Cuchillo-Ib´ a˜ nez b , c , Oscar Ramírez-Molina a , Tiare Silva-Grecchi a , Jessica Panes-Fern´ andez a , Patricio Castro d , Javier S´ aez-Valero b , c , Jorge Fuentealba a , e , * a Laboratorio de Screening de Compuestos Neuroactivos, Departamento de Fisiología, Facultad de Ciencias Biol´ ogicas, Universidad de Concepci´ on, Concepci´ on, Chile b Instituto de Neurociencias de Alicante, Universidad Miguel Hern´ andez-CSIC, Sant Joan d′Alacant, 03550 Alicante, Spain c Centro de Investigaci´ on Biom´ edica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), Spain d Departamento de Fisiología, Facultad de Ciencias Biol´ ogicas, Universidad de Concepci´ on, Concepci´ on, Chile e Centro de Investigaciones Avanzadas en Biomedicina (CIAB-UdeC), Universidad de Concepci´ on, Concepci´ on, Chile ARTICLE INFO Keywords: P2×2 Purinergic receptors Amyloid beta Alzheimer’s disease Fe65 APP ABSTRACT Amyloid beta peptide (Aβ) is tightly associated with the physiopathology of Alzheimer’s Disease (AD) as one of the most important factors in the evolution of the pathology. In this context, we previously reported that Aβ increases the expression of ionotropic purinergic receptor 2 (P2×2R). However, its role on the cellular and molecular Aβ toxicity is unknown, especially in human brain of AD patients. Using cellular and molecular approaches in hippocampal neurons, PC12 cells, and human brain samples of patients with AD, we evaluated the participation of P2×2R in the physiopathology of AD. Here, we reported that Aβ oligomers (Aβo) increased P2×2 levels in mice hippocampal neurons, and that this receptor increases at late Braak stages of AD patients. Aβo also increases the colocalization of APP with Rab5, an early endosomes marker, and decreased the nuclear/cytoplasmic ratio of Fe65 and PGC-1 α immunoreactivity. The overexpression in PC12 cells of P2×2a, but not P2×2b, replicated these changes in Fe65 and PGC-1 α ; however, both overexpressed isoforms increased levels of Aβ. Taken together, these data suggest that P2×2 is upregulated in AD and it could be a key potentiator of the physiopathology of Aβ. Our results point to a possible participation in a toxic cycle that increases Aβ production, Ca 2+ overload, and a decrease of PGC-1 α . These novel findings put the P2×2R as a key novel pharmacological target to develop new therapeutic strategies to treat Alzheimer’s Disease. 1. Introduction AD is characterized by two histopathological markers: neurofibrillary tangles and amyloid plaques [1], that are insoluble aggregates of amyloid-β peptide (Aβ). Dysregulation of APP processing and the subsequent increment in the Aβ levels are the key steps in AD pathogenesis [2]. Several evidence have proposed that soluble oligomer of Aβ (Aβo) are the main toxic species in AD that eventually lead to synaptic failure [3–8]. Aβ peptide is produced by the amyloidogenic cleavage of APP, this protein can also be proteolytically processed in the non-amyloidogenic pathway [9,10]. It has been described that these different processing pathways of APP occur in separate cellular compartments. The non-amyloidogenic processing mostly in the plasma membrane [11], while the amyloidogenic pathway takes place in Abbreviations: AD, Alzheimer’s disease; Aβ, Amyloid beta; Aβo, Oligomers of Aβ; AICD, Amyloid intracellular domain; APP, Amyloid precursor protein; ATP, Adenosine triphosphate; BDNF, Brain-derived neurotrophic factor; GSK-3β, Glycogen synthase kinase 3 beta; LRP1, Low-density lipoprotein receptor-related protein 1; LRP1b, Low-density lipoprotein receptor-related protein 1b; NMDA, N-methyl-D-aspartate; P2XR, Purinergic ionotropic receptor P2X; PPADS, Pyridoxalphosphate6-azophenyl-2 ′,4′-disulfonic acid; RME-6, Receptor-mediated endocytosis 6; sAPP α , Amyloid precursor protein soluble alpha; sAPPβ, Amyloid precursor protein soluble beta. * Correspondence to: Laboratorio de Screening de Compuestos Neuroactivos, Facultad de Ciencias Biol´ ogicas, Departamento de Fisiología, Universidad de Concepci´ on, PO Box 160-C, Barrio Universitario s/n, Concepci´ on, Chile. E-mail address: [email protected] (J. Fuentealba). 1 PA Godoy and D Mennickent should be considered joint first authors. Contents lists available at ScienceDirect Biomedicine & Pharmacotherapy journal homepage: www.elsevier.com/locate/biopha https://doi.org/10.1016/j.biopha.2021.111968 Received 13 April 2021; Received in revised form 20 July 2021; Accepted 23 July 2021 Biomedicine & Pharmacotherapy 142 (2021) 111968 2 intracellular compartments, mainly in endosomes [9,12]. This endocytosis of APP makes represent a key step in the production of the Aβ peptide [12,13]. It has been postulated that Aβo perforate plasma membranes [5,14], allowing the passage of small molecules and metabolites such as ATP, according to the concentration gradient [7,15]. This helps to explain some of the events described in models of AD that correlate with physiopathology features of the disease. I.e. Ca 2+ dyshomeostasis, ATP leakage, mitochondrial dysfunction and synaptic failure [15–19]. The increment in extracellular ATP induces a paracrine or autocrine activation of purinergic ionotropic receptors P2X (P2XR) in the vicinity. The 7 different subunits form trimeric cation permeable channels and act as neuromodulators of synaptic activity, inducing a facilitation of neurotransmitter release in neuronal and glial cells [20–22]. Some of these receptors are implicated in AD and other neurodegenerative diseases [23]. One of the toxic features observed in AD is chronic neuroinflammation [24], and the participation of P2×7R and P2×4R expressed in microglia has been observed [25]. Our group has shown the participation of P2XR in some of the toxic features induced by Aβo. One of this is the increment in intracellular Ca 2+ levels, that could be partially prevented by using PPADS, a P2XR antagonist [7], this suggests an increased P2X activation-ATP dependent. Then, we have described an increment in P2×2R in different cell cultures related to the chronic Aβo treatment [26]. P2×2R is one of the most widely distributed subtype of P2XR. It is abundantly expressed in the periphery and central nervious systems, a slow desensitization and high sensitivity to ATP (low half maximal effective concentration, EC 50 ) [27]. The presence of P2×2R is observed in axons and close to nerve terminals, where they participate in the regulation of neurotransmitter release in hippocampal interneurons [21]. One of the most important properties of these receptores is related to a relatively high calcium permeabilty [28], that could be important to maintain the cytosolic Ca 2+ homeostasis. The two main isoforms are P2×2a and P2×2b. P2×2a contains the complete exonic sequence, and P2×2b lacks 69 amino acids in the C terminus region [29]. This portion could be important to some relevant interaction with key intracellular factors, for example the interaction of P2×2a with Fe65 [29], an adaptor protein that can also interact with APP. The evidence shows that this interaction regulates APP endocytosis, and could regulate APP processing and Aβ generation [30]. Although P2×2R overexpression has been reported after treatment with Aβo, it is unclear the possible impact on the mechanisms of Aβo toxicity. Furthermore, no previous study has investigated the P2×2 expression on brain from patients with AD. In our present work, we explored the possible toxic effects of P2×2R overexpression and the crosstalk with Aβo. We evaluated the impact on some key intracellular proteins and mechanisms such as Fe65, APP endocytosis and Aβ generation, events that can induce a neuronal death and synaptic dysfunction. Furthermore, the finding of an altered presence of the receptor in the cortex of AD patient samples support our hypothesis that P2×2R is involved in the pathophysiology of AD. Together, these data represent a significant advance toward a comprehensive mechanism of Aβo toxicity and the participation of P2×2R in AD. 2. Materials and methods 2.1. PC12 cell culture PC12 cells were obtained from ATCC (Cat# CRL-1721, RRID: CVCL_0481, Manassas, VA, USA) and were cultured as described by Gavilan et al., [31]. Briefly, they were maintained in DMEM (Corning, NY, USA) supplemented with FBS (5%) (Gibco, Grand Island, NY, USA), HS (5%) (Gibco, Grand Island, NY, USA), and penicillin-streptomycin (1%) (Gibco, Grand Island, NY, USA). Cells were incubated under standard conditions (37ºC, 5% CO 2 ). When 70–80% confluence was achieved, the cells were treated or transfected with P2×2a (1 μ g), P2×2b (1 μ g) and/or mCherry (0.5 μ g) coding plasmids using Lipofectamine-2000 (Invitrogen, CA, USA). Cultures were used 24 h after transfection. 2.2. Mice primary hippocampal culture Pregnant C57BL/6 mice were treated in accordance with the regulations recommended by NIH and the Ethics Committee of the Universidad de Concepci´ on (Concepci´ on, Chile) complying with the current laws in Chile. Primary cultures of embryonic hippocampi (E18) were obtained as described in Gavilan et al., [31]. Briefly, cells were plated at 320,000 cells/ml on coverslips coated with poly-L-lysine (Trevigen, Gaithersburg, MD, USA) in MEM (Gibco, Grand Island, NY, USA) supplemented with 10% HS, 4 mg/ml DNAse and 2 mM L-glutamine (Gibco, Grand Island, NY, USA) for 24 h. Culture medium was replaced after 24 h with MEM, 2% HS, 2% FBS and 0.5% N 3 (BSA 1 mg/ml, putrescine 4 mg/ml, insulin 1.25 mg/ml, sodium selenite 1 µg/ml, TH3 2 µg/ml, progesterone 1.25 µg/ml, corticosterone 4 µg/ml). The cell cultures were maintained at 37ºC with 5% CO 2 . Experiments were performed at 10–11 DIV in control and treated neurons. 2.3. J20 hippocampal slices Brain slices were obtained from J20 transgenic mice, a C57BL/6 J mice with human mutated APP KM670/671NL (Swedish) and V717F (Indiana), and C57BL WT mice. Animals were manipulated in accordance with the ethical regulations established by NIH and University of Concepcion, and under the 3B criteria, one mouse from each phenotype was used to obtained the brain slices. Intracardiac perfusion was performed after 24 h of liquid diet and under Ketamine/Xylazine anesthesia. First, a perfusi´ on with saline solution was performed (NaCl 0,9%). Next, a fixing solution (paraformaldehyde 4%) was used. The brains were extracted and post-fixed in the same solution (4 h, 4 ◦C). Lastly, the brains were in etanol (24 h, 4 ◦C) and brain slices of 60 μ M were obtained in a vibratome Leica VT10005. 2.4. Collection of human brains This study was approved by the Ethics Committee of the Universidad Miguel Hern´ andez de Elche (Alicante, Spain) and was performed in accordance with the World Medical Association (WMA) Declaration of Helsinki. Brain samples (frontal cortex) were obtained from the Brain Bank at the Institute of Neuropathology, Bellvitge University Hospital (Barcelona, Spain). Cases of sporadic AD were selected on the basis of their clinical history of dementia and neuropathological CERAD diagnosis [32]. Samples were categorized according to the Braak Stage of neurofibrillary tangle pathology [33]. 12 women, 18 men; mean ±sd age, 75 ±10 yr; Braak stage I–II, n =10; Braak stage III–IV, n =10; and Braak stage V–VI, n =10, or alternatively by a rating of Aβ-phases [34]. Special care was taken not to include cases with combined pathologies to avoid bias in the pathologic series. Samples from non-demented (ND) controls (2 woman, 7 men; mean ±sd age, 53 ±5 yr) corresponded to individuals with no clinical dementia and no evidence of brain pathology. The mean postmortem interval of the tissue was approximately 8 h for all cases, with no significant difference between groups. 2.5. Protein extraction from human brain Frontal cortex (0.1 g) was homogenized (10% w/v) in Tris-HCl (50 mM, pH 7.4), NaCl (150 mM), Triton X-100 (0.5% w/v), Nonidet P-40 (1% w/v), and a cocktail of protease inhibitors (1:25 v/v, MilliporeSigma, Billerica, MA, USA). The homogenates were then sonicated on ice using an ultrasonic cell disruptor (Misonix, Farmingdale, NY, USA) and centrifuged for 1 h at 43,000 rpm and 4 ◦C. The supernatant fractions were recovered for further analyses by Western blot. P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 3 2.6. qRT-PCR analysis RNA was extracted from human brains using the TRIzol® Reagent (Thermo Fisher Scientific, MA, USA) in the PureLink™ Micro-to-Midi Total RNA Purification System (Life Technologies, CA, USA) following the manufacturer instructions. SuperScript™ III Reverse Transcriptase (Life Technologies, CA, USA) was used to synthesize cDNAs from total RNA (2 μ g) using random primers according to the manufacturer’s instructions. Quantitative PCR amplification was performed on a StepOne™ Real-Time PCR System (Applied Biosystems, CA, USA). Specific TaqMan probes for human P2×2 (assay ID: Hs04176268_g1, Thermo Fisher, MA, USA) and human GAPDH were used (Applied Biosystems, CA, USA). P2×2 was quantified using the relative standard curve method normalizing it by GAPDH from the same cDNA preparation. 2.7. Aβ 1–40 peptide Stock peptide (GenicBio, Shanghai, China) was reconstituted in DMSO to a concentration of 2.3 mM. To obtain Aβo, the peptide was aggregated in sterile H 2 O (Gibco, Grand Island, NY, USA) at 80 μ M using a standardized protocol (500 rpm, RT, 4 h). All treatments with Aβo were performed at a final concentration of 0.5 μ M in the culture medium of mice hippocampal cells for 24 h. The presence of Aβo has been previously tested [35]. 2.8. Electrophysiology The whole-cell patch clamp technique was performed using an Axopatch 200B amplifier (AxonInstruments, CA, USA) in voltage clamp mode at a holding potential of - 60 mV. The cells were exposed to ATP (1 mM) with a perfusion system. The pipette solution was (in mM): 120 KCl, 4 MgCl 2 , 10 HEPES, 2 ATP, 0.5 GTP, 10 BAPTA (pH 7.4, 300 mOsm). The bath solution contained (in mM): 150 NaCl, 5.4 KCl, 2 CaCl 2 , 1 MgCl 2 , 10 HEPES, 10 Glucose (pH 7.4, 320 mOsm) and TTX (50 nM). The analysis was performed using the Clampfit 10 software (AxonInstruments, CA, USA). 2.9. Ca +2 measurements Mice hippocampal cells were loaded with the non-ratiometric Ca +2 sensitive fluorescent probe Fluo-4 AM (Invitrogen, CA, USA) (5 μ M in DMSO) for 20 min in DPBS in standard incubation conditions. The cells were washed for 20 min with DPBS and finally washed 2 times with normal external solution, and then mounted on a Nikon TE-200-U inverted microscope (Tokyo, Japan). The cells were exposed to ATP (10 µM, 30 s) and high-potassium external solution (60 mM, 30 s) with a perfusion system. Changes in fluorescence were acquired using an iXon +EMCCD camera and analyzed with the Imaging Workbench 6.0 software (INDEC Biosystems, CA, USA). 2.10. Immunocytochemistry Cell cultures were fixed with 4% PFA for 15 min at 4 ◦C, then the cells were permeabilized and blocked (0.1% Triton X-100, 10% HS) for 30 min at RT. The samples were incubated with anti-MAP1B (1:200, Santa Cruz Biotechnology, Cat# sc-8970, RRID:AB_649156, goat), anti-P2×2 (1:200, Alomone Labs Cat# APR-025, RRID:AB_2341051, rabbit), antiAPP (1:200, Abcam Cat# ab32136, RRID:AB_2289606, rabbit), antiFe65 (1:50, Santa Cruz Biotechnology, Cat# sc-374641, RRID: AB_10987657, mouse), anti-Rab5 (1:200, Sigma-Aldrich Cat# R7904, RRID:AB_532319, mouse) and/or anti-PGC-1 α (1:200, Novus Cat# NBP1–04676, RRID:AB_1522118, rabbit) primary antibodies for 1 h at RT. Then they were incubated with the corresponding fluorescentlabeled secondary antibodies for 45 min (anti-rabbit Alexa Fluor 488, Jackson ImmunoResearch Labs Cat# 711–545–152, RRID:AB_2313584; anti-goat Cy3, Jackson ImmunoResearch Labs Cat# 305–165–003, RRID:AB_2339464; anti-goat Alexa Fluor 647, Jackson ImmunoResearch Labs Cat# 705–605–003, RRID:AB_2340436; anti-mouse Cy3, Jackson ImmunoResearch Labs Cat# 115–165–003, RRID:AB_2338680; and anti-mouse Alexa Fluor 647, Jackson ImmunoResearch Labs Cat# 715–606–151, RRID:AB_2340866). For nuclear staining, the cells were incubated with DAPI (300 nM, Tocris (Bristol, UK) for 10 min. The slides were mounted using Dako immunofluorescence mounting media (Dako, CA, USA). Images (60–63 ×) were acquired with Nikon TE-200-U inverted (Nikon, Tokyo, Japan), LSM780 NLO confocal (Zeiss, Oberkochen, Germany) and SIM ELYRA S.1 super-resolution (Zeiss, Oberkochen, Germany) microscopes. The analysis was performed using Image J (NIH, MD, USA). First, the images were deconvoluted, then for the quantification 5 different 10px x 10px ROIs were selected in each image using the marker protein in each experiment, to avoid bias in the region selection. The fluorescence intensity in these selected regions was graphed. 2.11. Immunohistochemistry Slices were washed five times with PBS (5 min per wash) and then incubated in blocking/permeabilization solution (0.3–0.5% Triton X100 +10% horse serum +PBS) for 1 h. Then, slices were incubated with the primary antibodies at ice cold temperature overnight at 4 ◦C. Antibodies used were P2×2 (rabbit, 1:400, Alomone Labs Cat# APR-02) and MA1B (mouse, 1:400, Santa Cruz) prepared in a solution containing 0.3% Triton X-100 +10% horse serum +PBS. Subsequently, slices were washed five times for 3 min each and incubated with the corresponding secondary antibody (conjugated with either Cy3 or AlexaFluor488) for 90 min. After incubation, nuclear staining was performed with DAPI 300 nM for 15 min. Afterwards, slices were washed five times for 3 min each and mounted with DAKO fluorescent mounting media. Images were acquired on a LSM780 NLO Zeiss confocal microscope using a 10X objective. Images were processed using both ZEN software (Carl Zeiss MicroImaging GmbH) and Image J (NIH, USA). Immunoreactivity was quantified with ImageJ using 5 regions of interest in each hippocampal area within each slice. 2.12. Western blot Lysates from PC12 cells, mice hippocampal cultures, and human brain extracts were denatured (37ºC, 30 min), subjected to SDS-PAGE (100 V, 100 min), and transferred to nitrocellulose membranes (250 mA, 120 min). They were blocked for 2 h with 5% non-fat milk in TBS-T or Odyssey blocking buffer (Li-COR, USA). Nitrocellulose membranes were cut to incubate with the different primary antibodies: Anti-P2×2 (1:500, Alomone; 1:2000, NBP2–19655 Novus, USA), anti-Aβ (1:1000, MOAB-2, NBP2–13075 Novus, USA), anti Gβ (1:1000, Santa Cruz Biotechnology Cat# sc-378, RRID:AB_631542) and anti GAPDH (1:3000, Abcam Cat# ab9485, RRID:AB_307275) ON at 4ºC. Secondary HRP-conjugated (1:5000, Santa Cruz Biotechnology Cat# sc-2313, RRID:AB_641181, Santa Cruz Biotechnology Cat# sc-2005, RRID: AB_631736) or IRDye-conjugated (1:8000, anti-rabbit IRDye 800CW, LICOR Biosciences Cat# 925–32213, RRID:AB_2715510 and anti-mouse IRDye 680, LI-COR Biosciences Cat# 925–68072, RRID:AB_2814912) antibodies were incubated for 1 h at RT. Immunoreactive bands were exposed using the Clarity Western ECL Substrate (Bio-Rad, CA, USA) and the Odyssey FC detection system (Li-COR, USA) or Odyssey CL×Infrared Imaging system (Li-COR, USA). Image analysis was done with the P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 4 ImageStudio software (Image Studio Inc., WI, USA). For the correct detection and quantification of the immunoreactive bands in the experiments with overexpression, more of the control lysates was loaded in the gel. Normalization was made for each band with the corresponding loading control protein. 2.13. Experimental Design and Statistical Analysis For treatments with Aβo mice primary hippocampal cultures were used, for overexpression of P2×2R the PC12 cell line was used. Human brain extracts were obtained from frontal cortex to evaluate P2×2 expression in patients with AD and control individuals. For immunocytochemistry samples at least three different areas were observed for each replicate. Different cells were objectively used to perform the immunorreactive analysis where random regions of interest (ROIs) were selected in each image to analyse the immunofluorescence intensity. Statistical analysis were performed in experiments with least three independent biological replicates. The only exception for this is the analysis in J20 brain slices, where we only used one animal following the principle of the 3 R. Statistical significance was determined using Student’s t-test and one-way ANOVA for normally distributed data, or Mann-Whitney U-test and Kruskal-Wallis H-test for not normally distributed data. GraphPadPrism 6.0 software (GraphPad Prism (CA, USA) was used for this anlysis. n corresponds to independent replicates, and N to the number of different cells analyzed. The following results were considered statistically significant: *p <0.05, * *p <0.01 vs control, #p <0.05, ##p <0.01 vs mCherry. In graphs the data is presented as the mean ±SEM. 3. Results 3.1. The chronic treatment of hippocampal neurons with Aβo induces an increment of P2×2R In our previous work we described that Aβo induced an increment in P2×2R expression levels in rat hippocampal neurons [26]. Now, we wanted to corroborate this observation in other cellular models such as mice hippocampal neurons, and to explore a more physiological approach such as hippocampal slices from the transgenic J20 mice. This mice expresses two mutant forms of the human APP that increases the levels of Aβ in their brains, a widely used model for AD studies [36]. To achieve this goal, we measured P2×2 protein levels in mice hippocampal cultures after a 24 h treatment with Aβo (0.5 µM, Fig. 1a). We found an increase of 30 ±11% with respect to the control conditions (Fig. 1b, n =5,). In parallel, we used an additional experimental approach to corroborate the observations in western blot. Using immunofluorescence and confocal microscopy we evaluated the P2×2 protein immunoreactivity (Fig. 1c), under the same experimental conditions. Quantification of the fluorescence intensity in neurons showed, in accordance with the western blot results, an increase of 33 ±12% in P2×2 immunoreactivity after Aβo exposure compared to non-treated cells (Fig. 1d, n =5). In another additional experimental approach, focused on provide complementary evidence about these observations, we studied the P2×2R on J20 transgenic mice. To carry out this, we used Fig. 1. P2×2 increases after Aβo treatment in mice hippocampal cells (a) Representative western blots for P2×2 and β-actin in mice hippocampal cells in control conditions and treatment with Aβo (0.5 μ M, 24 h). (b) Graph shows the quantification for P2×2 normalized to β-actin and presented as percentage of control (C: 100 ±6%, Aβo: 130 ±11%,). All values represent mean ±SEM, n =5, *p =0.0328, Two-tailed t-test (c) Immunofluorescence for MAP1B (red) and P2×2 (green) in mice hippocampal cells in control and treatment with Aβo (0.5 μ M, 24 h). (d) Graph shows the quantification for P2×2 immunoreactivity presented as percentage of control (C: 100 ±5%, Aβo: 133 ±12%). All values represent mean ±SEM, n =5, *p <0.0499, Mann-Whitney test. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article) P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 5 (caption on next page) P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 6 hippocampal slices from a J20 transgenic mice (10 months old) and we compared the immunoreactivity of P2×2R with a WT mice (10 month) in all key areas of hippocampus (CA1 Fig. 2a, CA3 Fig. 2c, Dentate Gyrus Fig. 2e). We quantified the P2×2R immunoreactivity in the body cells of neurons in these three key areas (Fig. 2b, d and f), and we observed a significant increase in CA1 and DG of 60 ±7% and 61 ±26%, respectively. This important observation helps us to suggest that the Aβo toxicity could be enhancing the purinergic tone mediated by an increment on P2×2R levels. To check if these observations are important in human Aβ physiopathology, the next step was to evaluate the levels of P2×2R in brain human postmortem samples. 3.2. P2×2 expression is upregulated in the cortex of patients with Alzheimer’s disease Several studies have attempted to elucidate key elements (NMDA receptor, BDNF, GSK-3β [37–39] that could be altered in AD patients as pathognomonic signals to develop therapeutic strategies, unfortunately without success to date. Here, we evaluated the expression of P2×2R in cortex samples of patients at different stages of AD and compared the data to control samples of non-demented individuals (ND). First, we studied P2×2 protein levels through Western blot using GAPDH as a loading control (Fig. 3a, b). We observed a significant increment of 54 ±15% in its levels in V-VI Braak stages compared to the ND controls (Fig. 3c, n =9 (ND) and 10 (Braak Stages). We noticed that in each Braak stage group, the samples that had the higher Aβ deposition staging had the higher expression of P2×2 (0: green, A: yellow, B: orange, and C: red). Hence, we quantified P2×2 in each Aβ deposition staging-organized-sample group (Fig. 3d). The analysis showed that the C stage presented a 53 ±14% increase compared to the ND samples (n =9 (ND), 8 (0), 9 (A and C9 and 4 (B)). To reinforce these observations, we evaluated the P2×2 mRNA expression in the same samples. We observed a non-significant increase in the levels of P2×2 mRNA in all AD Braak stages with respect to the control samples (Fig. 3e, n =8 (ND and I-II), 9 (III-IV) and 10 (V-VI)). Although, the samples presented a tendency to be increased according to the Aβ deposition staging. We quantified a significant increment of 132 ±50% in the A stage compared to the ND samples (Fig. 3f, n =8 (ND), 6 (0), 8 (A), 4 (B) and 9 (C)). These important results reinforce our hypothesis that P2×2R have a role in the physiopathology of AD. However, to consolidate this information and to demonstrate their effective contribution to Aβ toxicity, the overexpression of the receptor has to be functional. 3.3. The P2×2R overexpressed by Aβo is functional Considering our results, one of the most important aspects to propose a role for P2×2R in this toxicity model, was to check if the increment in P2×2 after Aβo exposure was functional. To evaluate this, first we used electrophysiological patch clamp approaches. We compares the ionic conductance of ATP-evoked currents (1 mM) in control conditions and after Aβo treatment (0.5 μ M, 24 h, Fig. 4a). We observed that the ATPevoked currents in neurons exposed chronically to Aβo were significantly higher (231 ±70%, Fig. 4b) compared to control. This indicates that the treatment overexpressed receptors (that we described in previous results) were functional, therefore these cells experiment an increment of ionic current mediated by P2×2R. To reinforce this evidence and taking into consideration that cytosolic Ca 2+ dyshomeostasis is a central element on Aβ toxicity and P2×2R are highly permeable to Ca 2+ , we measured intracellular calcium changes evoked by ATP (10 µM, 30 s, Fig. 4c). These experiments reported a higher Ca 2+ cytosolic signal in Aβo treated hippocampal neurons (194 ±24%, Fig. 4d) compared to control. To check that result was not related to a change on general cellular excitability, we used a global depolarizing experimental stimulus; high K + solution (60 mM, 30 s). Additionally, we used this maximal cellular response to normalize the Ca 2+ -ATP response on treated and non-treated neurons. Both experimental conditions presented the same K + -response (Fig. 4e). This result corroborates that the treatment with Aβo, does not affect the general excitability of the neurons and suggests that the increased response to ATP was directly related to an increase in the purinergic signaling. 3.4. Chronic treatment with Aβo decreases Fe65 nuclear/cytoplasmic ratio and PGC-1 α levels in mice hippocampal neurons After we studied P2×2R levels and function, we aimed to study Fe65 protein, a scaffold protein that interacts with both P2×2 [29] and APP [30]. First, we wanted to evaluate if the overexpressed P2×2R, induced by chronic treatment with Aβo, could induce changes in Fe65 and APP immunoreactivity. We used immunofluorescence techniques and confocal microscopy in hippocampal neurons (Fig. 5a), where we observed no changes in the Fe65 under the described experimental conditions (Fig. 5b). However, APP immunoreactivity increased by 47 ±16% in treated hippocampal neurons (Fig. 5c). Considering that Fe65 translocates to the nucleus to activate the transcription of key genes [40], we decided to study its distribution between the nucleus (using DAPI staining) and the cytoplasm (using MAP1B immunoreactivity, Fig. 5d). As shown in the graph of Fig. 5e, the N/C ratio of Fe65 immunoreactivity decreased 28 ±4% after chronic Aβo exposure compared to the control condition. This suggests that a reduced presence of Fe65 in the nucleus, could alter its function as a transcription regulator, compromising any reparative or corrective mechanisms to revert the Aβ insults. One of the key proteins whose gene expression is controlled by Fe65, is PGC-1 α [41], a transcriptional co-activator and a major regulator of mitochondrial biogenesis [42]. Therefore, we expected to find the total PGC-1 α immunofluorescence to be affected, in correlation with previous findings. Consequently, after Aβo treatment (Fig. 5f), PGC-1 α immunoreactivity decreased 15 ±5% compared to the control mice hippocampal neurons in epifluorescence images (Fig. 5g). These results suggest that the increased P2×2R, induced by Aβo exposure, may affect the distribution of Fe65, reducing its presence in the nucleus, and causing an impact on PGC-1 α levels. This could represent a set of cellular events that could be a point of “no return” in the neuron, where it is unable to support its metabolic and synaptic activity demands. Fig. 2. Increased P2×2 immunoreactivity in hippocampus from J20 transgenic mice (a) Immunofluorescence for P2×2 (green) and MAP1B (red) in hippocampal CA1 region from WT and J20 transgenic mice. Red squares represent how the quantification was performed in these images. (b) Graph shows the quantification for P2×2 immunoreactivity in hippocampal CA1 body cells, presented as percentage of control (C: 100 ±6%, Aβo: 160 ±7%). All values represent mean ±SEM, N =5, n=1 * *p <0.001, Mann-Whitney test. (c) Immunofluorescence for P2×2 (green) and MAP1B (red) in hippocampal CA3 region from WT and J20 transgenic mice. (d) Graph shows the quantification for P2×2 immunoreactivity in body cells presented as percentage of control (C: 100 ±11%, Aβo: 119 ±25%). All values represent mean ±SEM, N =5, n =1, ns, Mann-Whitney test. (e) Immunofluorescence for P2×2 (green) and MAP1B (red) in hippocampal Dentate Gyrus region from WT and J20 transgenic mice. (f) Graph shows the quantification for P2×2 immunoreactivity in hippocampal GD body cells, presented as percentage of control (C: 100 ±8%, Aβo: 161 ±26%). All values represent mean ±SEM, N =5, n =1 * *p <0.001, Mann-Whitney test. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article) P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 7 Fig. 3. Increased P2×2 expression in AD (a) Representative western blots for P2×2 and GAPDH in ND and different Braak stages of AD cortex samples. (b) Representative western blots for P2×2 and GAPDH in ND and V-VI brak stage of AD cortex samples. (c) Graph shows the quantification for P2×2 normalized to GAPDH and presented as percentage of control (ND: 106 ±11%, I-II: 125 ±17%, III-IV: 106 ±12%, V-VI: 154 ±15%). Different Aβ stages are shown by a color system 0: green, A: yellow, B: orange, C: red. All values represent mean ±SEM, n =9 (ND) and 10 (Braak Stages), *p =0.0202, Student’s t-test Braak V-VI compared to ND (d) Graph shows the quantification for P2×2 in the samples at different Aβ deposition stages normalized to GAPDH and presented as percentage of control (ND: 106 ±11%, 0: 102 ±17%, A: 125 ±18%, B: 132 ±19%, C: 153 ±14%). All values represent mean ±SEM, n =9 (ND), 8 (0), 9 (A and C) and 4 (B), *p =0.0194, Student’s t-test Stage C compared to ND. (e) Graph shows the quantification for P2×2 mRNA normalized to GAPDH and presented as percentage of control (ND: 100 ±26%, I-II: 203 ±54%, III-IV: 202 ±60%, V-VI: 209 ±46%). Different Aβ stages are shown by a color system 0: green, A: yellow, B: orange, C: red. All values represent mean ±SEM, n =8 (ND and I-II), 9 (III-IV) and 10 (V-VI), not significant p =0.3712, one-way ANOVA. (f) Graph shows the quantification for P2×2 mRNA in the samples at different Aβ phases normalized to GAPDH and presented as percentage of control (ND: 100 ±25%, 0: 162 ±57%, A: 232 ±50%, B: 138 ±72%, C: 240 ±61%). All values represent mean ±SEM, n =8 (ND), 6 (0), 8 (A), 4 (B) and 9 (C) *p =0.0348, Student’s t-test Stage A compared to ND. P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 8 Fig. 4. ATP response increases in hippocampal neurons after Aβo exposure (a) Representative ATP (1 mM, 3 s) evoked currents in mice hippocampal cells in control and treatment with Aβo (0.5 μ M, 24 h). (b) Graph shows the quantification for the I max presented as percentage of control (C: 100 ±50%, Aβo: 231 ±70%). All values represent mean ±SEM, n =9, * *p <0.0078, Mann-Whitney test. (c) Representative intracellular calcium traces evoked by ATP (10 μ M, 30 s) and K + (60 mM, 30 s) in mice hippocampal cells in control (blue) and treatment with Aβo (red, 0.5 μ M, 24 h). (d) Graph shows the quantification of the fluorescence in relative units (RFU) evoked by ATP normalized to the response to K + and presented as percentage of control (C: 100 ±12%, Aβo: 194 ±94%). All values represent mean ±SEM, n =5, * *p =0.0037, Mann-Whitney test. (e) Graph shows the quantification of intracellular calcium changes evoked by K + 60 mM exposure, no significant differences were found between controls or Aβo treated neurons. (C: 100 ±5%, Aβo: 92 ±7%). All values represent mean ±SEM, n =5, not significant, Mann-Whitney test. P.A. Godoy et al. Biomedicine & Pharmacotherapy 142 (2021) 111968 9 (caption on next page) P.A. Godoy et al.