Adenosine signalling to astrocytes coordinates brain metabolism and function – Nature

    Adenosine signalling to astrocytes coordinates brain metabolism and function – Nature

    All animal studies were performed in accordance with the European Commission Directive 2010/63/EU (European Convention for the Protection of Vertebrate Animals used for Experimental and Other Scientific Purposes) and the UK Home Office Animals (Scientific Procedures) Act (1986) with project approval from the Institutional Animal Care and Use Committee of the University College London. The animals were group-housed and maintained on a 12-h light cycle and had ad libitum access to water and food. The mice were housed at 24 °C ambient temperature with relative humidity kept at 60 ± 5%. The rats were housed at 22 °C ambient temperature and 55 ± 10% relative humidity.

    Hippocampal slice preparation

    Male and female wild-type and Adora2bflox/flox mice, transduced to express the improved Cre (iCre) recombinase or tdTomato in astrocytes or young Sprague–Dawley rats (postnatal day 21 (P21)–P30) were terminally anaesthetized with isoflurane, the brains were removed and hippocampal slices (300–350 μm) were cut in an ice-cold slicing solution containing: 64 mM NaCl, 2.5 mM KCl, 1.25 mM NaH2PO4, 0.5 mM CaCl2, 7 mM MgCl2, 25 mM NaHCO3, 10 mM glucose and 120 mM sucrose, saturated with 95% O2 and 5% CO2 (pH 7.4). Slices were then left to recover for 1 h in artificial cerebrospinal fluid (aCSF) containing: 119 mM NaCl, 10 mM glucose, 3 mM KCl, 2 mM MgSO4, 1.25 mM NaH2PO4, 26 mM NaHCO3 and 2 mM CaCl2 (pH 7.4; 300–310 mOsm).

    Organotypic slice preparation

    Organotypic hippocampal slice preparations were obtained from the brain of rat (P5–P7 of either sex) or mouse (P8–P10 of either sex) pups as previously described42. The animals were terminally anaesthetized with isoflurane, and the brains were isolated and placed in ice-cold Hanks’ balanced salt solution (HBSS; Thermo Fisher) without Ca2+, with added 20 mM glucose (total 25.6 mM glucose), 10 mM MgCl2, 1 mM HEPES, 1 mM kynurenic acid, 0.005% phenol red, 100 U ml−1 penicillin and 0.1 mg ml−1 streptomycin. Coronal brain slices (350 μm) were cut and plated on 0.4-μm membrane inserts (Millicell CM, Millipore). The slices were incubated in a medium containing 50% Opti-MEM-1 (Thermo Fisher), 25% FBS (Sigma-Aldrich), 21.5% HBSS, 25 mM glucose, 100 U ml−1 penicillin and 0.1 mg ml−1 streptomycin. After 3 days, the incubation medium was replaced with a fresh medium and subsequently replaced twice a week. Astrocytes were transduced to express the genetically encoded fluorescent cAMP sensor Epac-SH187 (ref. 14) or the PKA activity sensor AKAR4 (ref. 15), under the control of the GFAP promoter43. Adenoviral vectors AVV-Gfap-Epac-SH187 (1.0 × 1010 PFU ml−1) or AVV-Gfap-AKAR4 (1.0 × 1010 PFU ml−1) were added to the incubation medium after 9–14 days of incubation and the slices were used in the experiments 3–5 days after the transfection.

    Primary astrocyte cultures

    Primary astrocyte cultures were prepared from the brains of rat and mouse pups (P2–P3 of either sex) as previously described44. The animals were terminally anaesthetized by isoflurane, and the brains were removed and the brain regions of interest separated by dissection. After isolation, the cells were plated on poly-d-lysine-coated coverslips and maintained in DMEM medium (Thermo Fisher) with 10% FBS, penicillin (100 U ml−1) and streptomycin (0.1 mg ml−1) at 37 °C in a humidified atmosphere of 5% CO2 and 95% air for a minimum of 10 days before the experiments.

    Genetic targeting of astrocytes to express fluorescent sensors

    Widespread expression of fluorescent sensors in the forebrain was achieved following microinjections of viral vectors in neonatal mice (P0–P2 of both sexes)42. The pups were prepared for aseptic surgery, and the solution containing the viral vector was administered into a lateral cerebral ventricle. The microinjections (volume 1–1.5 μl per side) were made 0.25 mm lateral to the sagittal suture, 0.50–0.75 mm rostral to the coronal suture and 2 mm ventral from the surface of the skull. Pups were kept in groups of litters and returned to their mothers in their home cages after the injections. Imaging experiments were performed 3–4 weeks after the injections.

    Genetic targeting of hippocampal astrocytes to express iCre recombinase

    To knock down A2B receptor expression, hippocampal astrocytes of mice with floxed Adora2b gene (Adora2bflox/flox; P0–P2 or 3–5-month-old of both sexes)26 were transduced to express iCre recombinase using the adeno-associated viral vector AAV5-Gfap-eGFP-iCre (VB1131, Vector Biolabs). Transduction of astrocytes with the viral vector AAV5-Gfap-tdTomato (University of Pennsylvania Vector Core) was used as a control. Neonatal mice were injected with viral vectors as described above. To transduce astrocytes of adult mice, the animals were anaesthetized with isoflurane (5% induction, 2–3% maintenance, in O2-enriched air). Adequate depth of surgical anaesthesia was maintained and confirmed by the absence of a withdrawal response to a paw pinch. With the head of the animal secured in a stereotaxic frame, a midline dorsal incision was made to expose the surface of the skull. A small craniotomy was performed and hippocampal CA1 regions were targeted with one microinjection per side of either AAV5-Gfap-eGFP-iCre vector or AAV5-Gfap-tdTomato vector. Microinjections (volume 0.3–0.5 μl given at a rate of 0.1 μl min−1) were made 2.0 mm rostral, 1.5 mm lateral and 1.5 mm ventral from bregma. After the microinjections, the wound was sutured. For post-surgical analgesia, the animals received buprenorphine (0.5 mg kg−1, subcutaneously). No complications were observed after the surgery, and the animals gained weight normally.

    Two-photon excitation imaging of changes in [cAMP] and PKA activity in astrocytes

    Optical recordings of changes in [cAMP] and PKA activity in hippocampal astrocytes were performed in acute or organotypic brain slices placed in a custom-made flow-through recording chamber mounted on a stage of an Olympus FV1000 microscope or FemtoSmart imaging system, optically linked to a Ti:Sapphire MaiTai laser with λ2P = 820 nm (Spectra Physics), with the emission filters set for the detection of CFP and YFP fluorescence. Recordings were performed at approximately 33–35 °C in aCSF saturated with 95% O2 and 5% CO2 (pH 7.4). For timelapse recordings of [cAMP] or PKA activity in astrocytes of the CA1 region of the hippocampus before and after the stimulation of Schaffer collateral fibres (described in detail below), images were collected with 512 × 512 pixel frames (frequency 1 Hz) using a water immersion ×25 Olympus objective (NA 1.05). Control optical recordings were performed using the same experimental settings without Schaffer collateral fibre stimulation applied. The laser power intensity was kept below 4 mW throughout the experiment.


    Electrophysiological experiments in acute hippocampal slices were performed as previously described42,44,45. The slices were placed in a recording chamber mounted on a stage of an Olympus BX51WI upright microscope (Olympus) equipped with a LUMPlanFL/IR 40 × 0.8 objective coupled to an infrared DIC imaging system and an Evolve 512 EMCCD camera (Photometrics). A source of fluorescent light was an X-Cite Intelli lamp (Lumen Dynamics). Wide-field fluorescence images were acquired using Micromanager 4.1 (ImageJ plugin) software and various digital zooms to visualize transduced astrocytes. Schaffer collateral fibres were stimulated using a concentric bipolar electrode (pulse width of 100 μs; amplitude of 20–300 μA, corresponding to approximately one-third of the saturating response). Synaptic responses were induced by trains of Schaffer collateral fibre stimulations consisting of five pulses applied at 20 Hz and delivered 50 ms apart. Evoked field excitatory postsynaptic potentials (fEPSPs) were recorded using glass electrodes (1–2 MΩ) placed at a distance of more than 200 μm from the stimulating electrode in the CA1 region showing strong astrocytic expression of transgenes. In a typical experiment, the evoked fEPSPs were recorded for at least 60 min.

    Synaptic long-term potentiation (LTP) in the CA3–CA1 pathway was induced by high-frequency stimulation (HFS) of Schaffer collateral fibres32,42,45. Basal synaptic transmission was first tested by low-frequency Schaffer collateral fibre stimulation (trains of 5 pulses applied at 20 Hz every 30 s) and monitored with recordings of evoked fEPSPs for 15–20 min. The HFS was then applied to induce LTP — a protocol consisting of three trains of stimuli (100 pulses at 100 Hz), applied with 60-s intervals. The fEPSPs were recorded for 60–90 min after the induction of LTP. In these experiments, picrotoxin (100 μM) and CGP-52432 (5 μM) were added to the bath solution to block the inhibitory transmission. The recorded signal was amplified (Multipatch 700B) and processed using pClamp 10.2 software (Molecular Devices). Recordings were filtered and digitized; the fEPSP slope was measured for the first evoked response in each pulse train.

    Biosensor recordings of lactate and adenosine release

    Lactate and adenosine were recorded in acute brain slices using amperometric enzymatic microelectrode biosensors (Sarissa Biomedical)46,47,48. The sensors were placed in direct contact with the surface of the slice placed on an elevated grid in a flow chamber at 35 °C. A dual recording configuration of a null sensor (lacking enzymes) and lactate or adenosine biosensor was used, as previously described46,47. The null sensor was used to determine whether any nonspecific electroactive interferents were detected and confounded the measurements. Null sensor currents were subtracted from the lactate or adenosine biosensor currents, and the resulting current profile was used to calculate the amount of the released analyte. In some experiments, adenosine and lactate signals were recorded simultaneously. Biosensors were calibrated with a known amount of lactate or adenosine added to the perfusate flowing through the recording chamber (in the identical temperature, aCSF composition and osmolarity conditions) immediately before and after the recordings. The mean of the initial and final calibrations was used to convert changes in sensor current to changes in lactate or adenosine concentration.

    Recordings of [cAMP], PKA activity, [glucose] and NADH–NAD+ redox state in cultured astrocytes

    Primary cultured astrocytes were transduced using viral vectors to express genetically encoded fluorescent sensors of cAMP (Epac-SH187 (ref. 14)), PKA activity (AKAR4 (ref. 15)), the cytosolic NADH–NAD+ redox state (Peredox28) or glucose (FLIP12glu-700μΔ6 (ref. 27)). The cells were incubated with a viral vector for 12 h and used in the experiments after 3 days following transduction. Optical recordings of changes in [cAMP], PKA activity, [glucose] and NADH–NAD+ redox state in astrocytes were performed using an inverted wide-field Olympus microscope equipped with a ×20 oil immersion objective lens, a cooled CCD camera (Clara, Andor, Oxford Instruments), a Xenon arc lamp, a monochromator and an Optosplit (Cairn Research). Recordings were performed in a custom-made flow-through chamber at 32–34 °C in aCSF saturated with 95% O2 and 5% CO2 (pH 7.4). The rate of chamber perfusion with aCSF was 1 ml min−1. To record FRET signal changes (cAMP, PKA activity and glucose sensors), 415/10 nm excitation light was applied, and the fluorescence emission was recorded at 470/24 and 535/30 nm. To record changes in the cytosolic NADH–NAD+ redox state, the Peredox sensor was excited with 405/10 nm and 575/10 nm light and the fluorescence emission was recorded at 535/30 nm and 630/35 nm.

    Conditional deletion of Adora2b in astrocytes

    To induce conditional A2B receptor knockdown in brain astrocytes, mice carrying a loxP-flanked Adora2b allele (Adora2bflox/flox)26 were crossed with the mice expressing an inducible form of Cre (Cre/ERT2) under the control of the astrocyte-specific Aldh1l1 promoter29. Recombination specificity of Aldh1l1Cre/ERT2 mice has been previously described29. Breeding was organized through PCR genotyping obtained from ear DNA biopsies. Tamoxifen (100 mg kg−1 dissolved in corn oil; injected intraperitoneally (i.p.) daily for 5 consecutive days) was given to Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cre/ERT{{2}}^{+}}\) and Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cr{e}^{-}}\) mice at 12–16 weeks of age. In separate groups of Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cre/ERT{{2}}^{+}}\) and Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cr{e}^{-}}\) animals, corn oil was given as a vehicle control. The expression of the A2B receptor in the brain was examined 4 weeks after tamoxifen treatment.

    The specificity of genomic recombination of the Adora2b locus was evaluated by PCR of the brain tissue of Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cre/ERT{{2}}^{+}}\) and Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cr{e}^{-}}\) mice treated with tamoxifen. The following primers were used to identify the 719-bp-long recombination product: recombination forward 5′-CAGTGCTGAGGCTATTAAAAAGGG-3′ and recombination reverse 5′-GGTGACTGCATAGCCTAGGGAAAC-3′. For PCR genotyping, the following primers were used: Adora2bflox/flox forward: 5′-TTAAAAGGTGATTCCCAGCACG-3′; Adora2bflox/flox reverse: 5′-GGTGACTGCATAGCCTAGGGAAAC-3′; Aldh1l1Cre/ERT2 forward: 5′-CTTCAACAGGTGCCTTCCA-3′; Aldh1l1Cre/ERT2 reverse: 5′-GGCAAACGGACAGAAGCA-3′. No Adora2b recombination was observed in tissues of control animals.

    Isolation of astrocytes and A2B receptor protein quantification

    Hippocampal astrocytes were isolated from the brains of Adora2bflox/flox or wild-type mice injected with AAV5-Gfap-eGFP-iCre vector to determine the effectiveness of A2B receptor deletion in this model. The animals were euthanized by isoflurane overdose, perfused transcardially with ice-cold saline and the hippocampal regions were isolated. The tissue was enzymatically dissociated to obtain a suspension of individual cells. Astrocytes were identified and sorted by eGFP expression (ARIA II BD). The purified fraction of eGFP-expressing astrocytes was lysed by sonication at low frequency using a Soniprep 150 Sonicator (Sanyo). Supernatant was collected after centrifugation and used for quantification of A2B receptor protein by ELISA (E03A1281, BluGene Biotech). A2B receptor protein quantification was normalized to the number of eGFP-positive cells in each sample.


    At the end of the experiments, Adora2bflox/flox mice transduced to express iCre recombinase or tdTomato in hippocampal astrocytes were given an anaesthetic overdose (sodium pentobarbital; 200 mg kg−1; i.p.), the brains were removed, fixed in 4% paraformaldehyde for 12 h and sliced (20 μm). Identification of eGFP expression in the brains of mice transduced to express iCre recombinase in hippocampal astrocytes was aided by antibody labelling. After slicing, free-floating sections were incubated with chicken with anti-GFP antibody (1:500; anti-GFP1020, Aves Labs or AB13970, Abcam) for 12 h at 4 °C. Sections were then incubated with secondary anti-chicken antibody AlexaFluor 568 (1:200; A-11041, Thermo Fisher) or AlexaFluor 488 (1:1,000; A-21441, Thermo Fisher) for 2 h at room temperature. Brain sections were mounted onto slides with Fluoroshield with DAPI (Sigma). Tiled images of coronal cross-sections were obtained using a Zeiss 800 confocal microscope.

    To evaluate the cell specificity of transduction with the AAV5-Gfap-eGFP-iCre vector, astrocytes, oligodendrocytes, neurons and microglia were labelled using the following antibodies: rabbit anti-GFAP (1:500; 23935-1-AP, Proteintech), rabbit anti-MBP (1:200; MA5-35074, Thermo Fisher), rabbit anti-NeuN (1:200; AB236870, Abcam), and rabbit anti-Iba1 (1:200; GTX100042, GeneTex), respectively. Secondary anti-rabbit antibody AlexaFluor 568 (1:1,000; 175470, Abcam) was used to identify the transduced cells.

    Western blot

    Astrocyte cultures prepared from Adora2Bflox/flox mice, transduced with either AAV5-Gfap-eGFP-iCre or AAV5-Gfap-tdTomato, were washed twice with PBS and the cells were collected in ice-cold lysis buffer supplemented with protease and phosphatase inhibitors (Thermo Fisher). Samples were snap frozen, sonicated and centrifuged at 14,000 rpm; the protein content of the extracts was determined by the Pierce BCA protein assay (Thermo Fisher). Twenty-five micrograms of protein was then fractionated on a Mini-PROTEAN TGX stain-free polyacrylamide gel (10%) (Bio-Rad) under denaturing and reducing conditions. Total protein levels were visualized in the gel after 3 min of exposure to UV light using a UV-transilluminator. Proteins were then transferred to polyvinylidene fluoride (PVDF) membranes (Bio-Rad), which were incubated with 5% non-fat milk. Membranes were next incubated overnight with the primary antibodies diluted in 5% BSA: rabbit anti-A2BR antibody (4 μg ml−1; AB1589P, Merck Millipore) and mouse anti-actin antibody (1:5,000; 3700, Cell Signaling Technologies), followed by incubation with the corresponding species-specific horseradish peroxidase (HRP)-conjugated secondary antibodies (1:5,000; anti-rabbit-HRP, sc-2054, Santa Cruz and anti-mouse-HRP, sc-2005, Santa Cruz). The luminol-based Pierce ECL Western Blotting Substrate (Thermo Fisher) was used to detect the HRP activity. After scanning the X-ray films, protein band densities were quantified using ImageJ.

    Quantitative real-time PCR

    Quantitative real-time PCR (rt–qPCR) assay was used to determine the level of Adora2b expression in the hippocampal tissue of Adora2Bflox/flox mice transduced to express iCre recombinase or tdTomato in astrocytes and in Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cre/ERT{{2}}^{+}}\) and Adora2bflox/flox:\(Aldh{1}l{{1}}^{Cr{e}^{-}}\) mice treated with tamoxifen. Total RNA was extracted, purified (RNeasy mini kit, 74106, Qiagen) and reverse transcribed using the QuantiTect Reverse Transcription Kit (205311, Qiagen) as per the manufacturer’s protocol. PCRs were performed in 20-μl volumes using the TaqMan Universal Master Mix II (4440040, Thermo Fisher) with a final volume of 9 μl cDNA, equivalent to 25 ng of RNA, sample template per reaction. PCRs were performed using the TaqMan assay (Adora2b, Mm00839292_m1, 61-bp amplicon length, Thermo Fisher) as detection method and an Agilent Technologies Aria Mx Real-time PCR system (Agilent). Adora2b expression was quantified using the comparative CT method (∆∆Ct) and presented as arbitrary units of expression, normalized to the expression of the ubiquitin C gene (Mm01201237_m1, 92-bp amplicon length, Thermo Fisher).

    Analysis of single-cell RNA-seq data

    Single-cell RNA-seq data of the mouse brain were obtained from a publicly available database collated and maintained by the Linnarsson group (Karolinska Institutet; Cell dissociation, single-cell RNA-seq and quality control methods are described in detail in the original report of the database24. Data processing and visualization were performed using the Seurat package49 in R (v.4.2.2, ‘Innocent and Trusting’). The combined mouse cortical cell RNA-seq dataset was obtained from 50,478 cells with expression data for 27,998 genes. All cells that displayed nFeatures greater than 200 and less than 4,000 and a percentage of mitochondrial RNA of less than 30% were included in the analysis. Remaining were 49,703 cells with average UMI counts (absolute number of observed transcripts; nCount) of 3,124.92 and nFeatures (genes per cell) of 1,592.39. The data were then log normalized and scaled to 10,000 transcripts per cell. The FindVariableFeatures function50 was used to identify the 4,000 most variable genes between the cells to be used in the principal component analysis (PCA). Before PCA, data were scaled with a linear transformation to ensure that all genes were given equal weight in the subsequent analyses. Dimensionality reduction was then performed by PCA on the scaled data up to and including the first 100 identified principal components. An elbow plot was used to determine the effective number of principal components, which was found to be 75. The k-nearest neighbour (KNN) graph was constructed using these 75 principal components. To cluster the cells, the Louvain method for community detection (Louvain algorithm) was used with resolution set to 2.0 as recommended for the datasets of this size51. Uniform manifold approximation and projection (UMAP) was used to visualize the cell clusters in two dimensions based on the same 75 principal components used for clustering and yielded 63 distinct cell clusters. The identity of cells comprising these clusters was determined by the differential expression of characteristic cell-specific marker genes24. The distribution of Adora2a and Adora2b expression was then plotted across the identified clusters. In addition, to determine the distribution of Adora2a and Adora2b expression restricted to the astrocyte-like cells in the sample, the expression data from cells identified as astrocytes in the initial projection were pooled and reclustered by the same method described for the whole dataset with modifications. The KNN graph was constructed using 20 principal components with Louvain algorithm resolution set to 0.8. Finally, the distribution of Adora2a and Adora2b expression was determined in all the identified cell-type clusters with more than 150 members where the cluster had at least 10% of all included cells that were found to be positive for either Adora2a or Adora2b.


    Mice were taken from their home cages, terminally anaesthetized with isoflurane overdose and transcardially perfused with ice-cold aCSF saturated with 95% O2 and 5% CO2. The brains were quickly isolated and snap-frozen in liquid nitrogen. The time taken from removing the animal from its habitat to the preparation of the frozen sample did not exceed 5 min in each case.

    Small molecules from the brain tissue were extracted as previously described52. In brief, frozen brain samples (100–150 mg) were transferred into 2-ml soft tissue homogenizing tubes (CK14, Precellys), kept on dry ice and 0.6 ml pre-chilled methanol:chloroform (2:1 v:v) solution was added to the samples. Samples were transferred to a bead beater (Precellys) and homogenized (within 1 min of removing from dry ice to keep metabolite profiles) two times for 10 s at 10,000 rounds per min. Subsequently, 0.2 ml of water and 0.2 ml of CHCl3 were added to each tube; the samples were vortex mixed and centrifuged at 13,000g for 10 min. The resulting top aqueous layer was aliquoted into microtubes, dried in Speedvac (Savant; 30 °C, overnight, VAQ setting) and then kept at −80 °C until assayed. A ‘pooled quality control’ sample was prepared by mixing 50 µl of the aqueous layer of each sample. The samples were reconstituted in 35 µl of water and analysed by ion-pairing liquid chromatography–mass spectrometry (LC–MS)53, using a XEVO TQ-S tandem mass spectrometer and an Acquity ultraperformance liquid chromatography binary solvent manager equipped with a CTC autosampler (Waters). Data were acquired with an electrospray ionization in a negative-ion mode and chromatography using a Waters HSS T3 column (1.8 µm, 2.1 × 100 mm) with a binary solvent system of 10 mM tributylamine + 15 mM acetic acid in water (as mobile phase A) and 80% methanol + 20% isopropanol (as mobile phase B) with a gradient elution. The sample processing order was randomized. Injections of double blanks (water) and single blanks were performed to ensure system stability, and to identify carryover and solvent interference peaks. The pooled quality control sample was injected at the beginning of the run and then once every tenth injection throughout the run, to monitor the instrument stability across the entire analytical session. The pooled quality control sample was used to evaluate the normalization method. During sample preparation and mass spectrometry analysis, the investigators were blinded to the identity of the experimental samples. The list of all the annotated metabolites and raw data are provided in Supplementary Table 1 and available at

    The LC–MS data were processed using Skyline54. Manually curated peaks were annotated based on in-house database of m/z and retention time of external standards; only peaks that passed the experimental ion ratios for the product ions (where more than one product existed) were integrated and peak area values were then exported. All data were normalized using the probabilistic quotient and analysed using R (v4.1.3). Relative levels of metabolites were mean centred and variance adjusted using the scale function of R. Multivariate classification models were built using partial least squares-discriminant analysis (PLS-DA) with the ropls R package (v.1.26.4) using sevenfold cross-validation for 10,000 permutations. Metabolomics pathway enrichment analysis was performed using MetaboAnalyst55 R package (v5.0) and the SMPDB database of metabolic pathways, with the top five metabolites differentiating between the experimental groups in the PLS-DA model as input.

    Novel object recognition test

    The experiments were performed in an isolated room under dim light conditions. Before testing, the animals were handled by the investigator daily for at least 1 week before the main experiment. The animals in their home cages were brought into the behavioural testing room 1 h before the experiment. First, each mouse was allowed to explore an empty square experimental chamber (40 × 40 × 40 cm) for 5 min. The animal was presented for 5 min with two identical objects placed diagonally on the floor of the chamber and then the animal was returned to the home cage. After 1 h, the animal was returned to the testing arena in which one of the original objects was replaced with a new object. The behaviour of the animal in the arena was recorded by tracking the nose of the mouse using a video camera and Viewer III software (Biobserve). The recognition memory was assessed by calculating the time the animal spent and the frequency of visits the animal made near each object during the training and testing sessions. The discrimination index (DI) as a measure of recognition memory was calculated using the formula DI = (time spent at novel object – time spent at familiar object)/(time spent at novel object + time spent at familiar object) × 100.

    EEG and EMG recordings

    For electrode implantation, the mice were anaesthetized with isoflurane (5% induction, 2–3% maintenance, in O2) and received buprenorphine (0.5 mg kg−1, subcutaneously) for perioperative analgesia. Adequate depth of surgical anaesthesia was maintained and confirmed by the absence of a withdrawal response to a paw pinch. With the head of the animal in a stereotaxic frame, a midline dorsal incision was made to expose the surface of the skull. EEG and EMG electrode headmounts (Pinnacle Technology) were secured to the skull using stainless steel screws and silver epoxy was used for optimal electrical connectivity. Two EMG leads were inserted into the nuchal muscles, and the headmounts were secured with dental acrylic. After a 10-day recovery period in a room with 12–12 light–dark cycle, mice were placed in individual Plexiglas circular recording cages (Pinnacle Technology) with unlimited access to water and food. The headmounts were connected to a lightweight EEG preamplifier (Pinnacle Technology) to enable unrestricted movement. Following a 3-day habituation period, EEG signals were sampled at 400 Hz using Sirenia software (Pinnacle Technology). Sleep stages were scored in 4-s epochs using SleepSign for Animal software (Kissei Comtec). Periods of wakefulness were identified by low-amplitude, high-frequency EEG and high EMG activity; NREM sleep was identified by high-amplitude, low-frequency EEG with minimal EMG modulation; and REM sleep was identified by low-amplitude, desynchronized EEG with low or absent EMG activity.

    Data analysis

    Imaging data were acquired using IQ3 software (v6.3; Andor, Oxford Instruments) or Olympus FluoView software (v4; Olympus) and analysed using Fiji (ImageJ). Biosensor recordings were acquired using Power 1401 interface and analysed using Spike2 software (v7; Cambridge Electronic Design). Electrophysiological data were acquired and analysed using pClamp software (v10.2). Statistical analysis of the data was performed using Origin 2019 software (v9.6) and GraphPad Prism software (v8). Distribution of data was analysed by a Shapiro–Wilk normality test. Grouped data were analysed using one-way or mixed-model ANOVA or Kruskal–Wallis test (for non-normally distributed data) when comparing data between more than two groups. One-way ANOVA was followed by Dunnett’s post-hoc test when comparing experimental groups against one control group or by Sidak’s post-hoc test when multiple comparisons between the groups were made. Comparisons of data obtained in the experiments with two groups were made using t-test or Mann–Whitney U-test (for non-normally distributed data). EEG data were analysed by repeated-measures ANOVA when multiple measurements were made over time in the same groups followed by Tukey’s post-hoc multiple comparisons test. The data are reported as individual values and mean ± s.e.m. or box-and-whisker plots. In the box-and-whisker plots, the central dot indicates the mean, the central line indicates the median, the box limits indicate the upper and lower quartiles, and the whiskers extend to 1.5 times the interquartile range from the quartiles. Details of the statistical tests applied are provided within the figure legends.

    Reporting summary

    Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

    Source link