Olfactory bulb astrocytes link social transmission of stress to cognitive adaptation in male mice

All animal protocols were in accordance with the Guidelines for the Animal Care and Use and the European Communities Council Directive of September 22th 2010 (2010/63/EU, 74) and approved by the French Ministry of Agriculture and Fisheries (authorization number 3306369) and the local ethical committee (authorization APAFIS#22372 and APAFIS #23685).

Animals

C57BL/6-N (Janvier, France) and inbreed constitutive and conditional CB1 mutant (center’s facility, with a predominant C57BL/6-N background) male mice (Mus musculus) were used for the different experiments of this project. CB1 mutant mice included: CB1^f/f mice (CB1-flox) carrying a floxed version of the CB1 gene³⁹; CB1-knockout mouse line (CB1-KO) carrying a constitutive global deletion of the CB1 gene³⁵; NEX-CB1-knockout mouse line (Glu-CB1-KO) carrying a conditional deletion of the CB1 gene in forebrain glutamatergic neurons under the control of a Nex-Cre recombinase³⁶; DLX-CB1 knockout mouse line (GABA-CB1-KO) carrying a conditional deletion of the CB1 gene under the control of a Dlx5/6-Cre recombinase³⁶; GFAP-Ert2-CB1-knockout mouse line (GFAP-CB1-KO) carrying an inducible conditional deletion of the CB1 gene in GFAP-expressing cells (mostly astrocytes) under the control of a GFAP-Cre recombinase²³ and a knock-in mouse line replacing the wild-type CB1 gene by a truncated form of the CB1 gene lacking the first 22 amino acids that reduces its mitochondrial-associated localization (DN22-CB1-RS)^26,38. The respective wild-type littermates of all lines were used as controls for the behavioral experiments.

Constitutive and inducible CB1 mutant mice were used in behavioral experiments. In the case of the GFAP-CB1-KO mice, they were injected with 8 daily injections of tamoxifen (Sigma, #T5648,1 mg, i.p.), dissolved in 90% sesame oil, 10% ethanol to a final concentration of 10 mg/ml to induce the CreERT2 dependent CB1 gene locus excision 4 weeks before the beginning of the behavioral experiments. CB1-flox mice were used for surgical procedures to specifically assess the role of CB1 in the olfactory bulb. C57BL/6-N mice were used as demonstrators (DEM) in all behavioral experiments using surgically induced mutant mice, and for surgical procedures to assess the role of mitochondrial Ca²⁺ in astrocytes in the olfactory bulb.

Non-littermates C57BL/6-N mice coming from outside the facility and facility inbreed mutant mice were housed together at 3 weeks of age (directly post weaning) in collective cages of 6-8 individuals. All animals were housed in the animal facility of Neurocentre Magendie with controlled temperature of 21 ± 2 °C, humidity 55%, in a 12 h light/12 h dark cycle (light on at 7.00am) and with water and food ad libitum. Animals were used at 8–17 weeks of age for the surgical and behavioral procedures, and assigned semi-randomly to experimental procedures (maintaining a balance between genotypes when required).

Adeno-associated viruses (AAV)

To generate a specific deletion on astrocytes of the OB, we used an AAV-hGFAP-Cre-IRES-mCherry purchased from the University of North Carolina (UNC School of Medicine) and an AAV-hGFAP-GFP or AAV-GFAP-dsRed as a control. To generate the specific deletion of neurons in the olfactory bulb, we used AAV-hSyn-Cre-GFP (Addgene catalog number #105540), and its control AAV-hSyn-GFP (Addgene, catalog number #105539). The AAV-CAG-Empty (used as control), AAV-CAG-DIO-CB1-GFP (expressing the wildtype CB1 construct) and AAV-CAG-DIO-DN22-GFP (expressing the DN22-CB1 construct excluding the mitochondrial associated location of the receptor) were used to specifically manipulate CB1 subcellular populations in vivo^24,38. The AAV-GFAP-mMICU1-S124A-HA-IRES-mRuby (expressing a mutated non-phosphorylable form of the MICU1 subunit of the mitochondrial Ca²⁺ transporter) and AAV-GFAP-mMICU1-WT-HA-IRES-mRuby (expressing the wildtype version of MICU1) were used to study the effects of mitochondrial calcium dynamics in vivo³⁴. The AAV-GFAP-mito-GcAMP6s and AAV-GFAP-GcAMP6f were used for fiber photometry experiments³⁴. The titrations of all viruses were between 10¹⁰ and 10¹¹ genomic copies per ml for all batches.

Surgery for viral injection and fiber implantation

Mice were injected intraperitoneally with burprenorphine (0.05 mg/kg, Buprecare), sleep-induced using 5% isofluorane, and placed into a stereotaxic apparatus (Model 900, Kopf instruments, CA, USA; with mouse adaptor and lateral ear bars) using 2% isofluorane for the duration of the surgery. Local analgesia with lidocaine (0.1 ml at 0.5%, Lidor) was used under the skin of the head before incision. The viral injections were delivered bilaterally in the olfactory bulb through a glass pipette using a microinjector (NanoInject II, Drummond Scientific). In all surgeries, mice were injected bilaterally with two injections per site of a total volume of 0.45 µl each in the following coordinates: AP + 4.1; ML ± 0.75; DV – 3 and − 2 at a speed of 5 nl/s.

To assess the specific contribution of astrocytic CB1 receptors in the olfactory bulb to socially-transmitted stress-driven behaviors, CB1-flox mice were injected with a viral mix of two different viruses: AAV-GFAP-GFP/AAV-DIO-Empty (expressing GFP reporter protein in astrocytes as a control, Ctrl), AAV-GFAP-CRE-mCherry/AAV-DIO-Empty (generating a Cre-induced deletion of CB1 receptors in GFAP positive cells, OB-GFAP-CB1-KO), AAV-GFAP-CRE-mCherry/AAV-DIO-CB1-GFP (generating both a Cre-mediated deletion of CB1 receptors in astrocytes and a Cre-mediated re-expression of the wild-type construct of CB1, OB-GFAP-CB1-RS) and AAV-GFAP-CRE-mCherry/AAV-DIO-DN22-GFP (generating both a Cre-mediated deletion of CB1 receptors and a Cre-mediated re-expression of the DN22-CB1 construct in astrocytes, therefore re-expressing CB1 everywhere but in their mitochondrial-associated locations, OB-GFAP-DN22-RS). All viruses used were titered between 2–8.10¹⁰ genomic copies/mL.

To exclude a neuronal contribution to impact on NOR of socially-transmitted stress, we injected CB1-flox mice with either an AAV-Syn-Cre-GFP or a AAV-Syn-EGFP (Ctrl) in the OB (titered 3,3.10¹¹ genomic copies/mL).

To measure in vivo Ca²⁺ calcium responses and the contribution to astrocytic CB1 receptors to this process, mice were injected in the OB with either only AAV-GFAP-mitoGcAMP6s (C57BL/6-N mice); AAV-GFAP-GcAMP6f (C57BL/6-N mice) or in combination (CB1flox mice) with AAV-GFAP-CRE-mCherry (OB-GFAP-CB1-KO), AAV-GFAP- dsRed (Ctrl). To generate the rescue and mitochondrial-specific mutants, we used mice coinjected with the GFAP-Cre and the DIO/Flex constructs AAV-DIO-CB1 (OB-GFAP-CB1-RS) or the AAV-DIO-DN22 (OB-GFAP-DN22-RS). All constructs were titered 2–5.10¹¹ genomic copies/mL. Then, the optical fiber (400 μm diameter, 0.5 NA) was placed 200 μm above the last injection site (at DV −2, therefore at −1.8) and fixed with dental cement (MajorRepair).

To assess the contribution of mitochondrial Ca²⁺ in socially-transmitted stress, C57BL/6-N mice were injected with either AAV-GFAP-MICU^WT or AAV-GFAP-MICU^S124A in the OB (titered 3.10¹¹ genomic copies/mL).

Following surgery, all mice received i.p. injection of 0.2 ml of saline solution and anti-inflammatory drug meloxicam (5 mg/kg, Metacam), that was continued for 2 additional days. Animals continued to be housed collectively and body weight was monitored daily during 4–5 days to assess recovery. Behavioral experiments were carried out 4–5 weeks after surgery and fiber photometry experiments 5–6 weeks after surgery.

Immunostaining For Light Microscopy

AAV injected mice were deeply anesthetized with pentobarbital (400 mg/kg body weight), transcardially perfused first with 20 ml of phosphate-buffered solution (PBS 0.1 M, pH 7.4) following by 30 ml of cold 4% paraformaldehyde (Sigma, BO501128-4L). Brains were isolated and postfixed in the same fixative solution overnight at 4 °C and then transferred to a 30% (wt/vol) sucrose (Sigma, S0389) solution in PBS for cryopreservation. Brains were then frozen in isopentane (Sigma, M32631) and stored at −80 °C. Free-floating frozen sagittal sections (30 µm) were cut using a cryostat (Leica Biosystems, CM1950S). Mid olfactory bulb slices were stored in antifreeze solution at −20 °C until further use.

Immunostaining against GFAP

Sections were washed with PBST (0.3% Triton X-100 diluted in PBS 1X pH7.4) three times and then permeabilized 1 h at room temperature (RT) in a blocking solution [in PBS 1X: 10% donkey serum; 0.3% triton X-100]. Next, sections were incubated overnight at 4 °C with polyclonal rabbit anti-GFAP (1:1000) (Agilent, DAKO Z0334) diluted in blocking solution. After washes with PBST, brain sections were incubated for 2 h at RT with donkey anti-rabbit Alexa fluor 647 (1:500, Invitrogen) (polyc). Following washes with PBST, sections were stained with DAPI (1:20000; Invitrogen D3571), washed again with PBST and finally mounted and coverslipped.

The sections were analyzed with an epifluorescence Leica DM6000 microscope (Leica, France) to check for the intrinsic fluorescence of the viruses and the identity of the infected cells. Mouse brains that did not meet the expression requirements led to the exclusion of the mice from the experiments.

Immunostaining against GFAP and HA

Sections were washed with PBST (0.3% Triton X-100 diluted in PBS 1X pH7.4) three times and then incubated with 3% hydrogen peroxide diluted in PBST (Sigma, H1009-500ML) for 30 min. Following a step of permeabilization carried out for 1 h at RT in a blocking solution, sections were incubated overnight at 4 °C with a mix of primary antibodies: polyclonal chicken anti-GFAP (1:1000) (USBiological #G2032-25F) and monoclonal rabbit anti-HA (1:1000, Cell Signaling, 3724) diluted in blocking solution. After some washes with PBST, brain sections were incubated for 2 h at RT with a mix of secondary antibodies: goat anti rabbit IgG HRP linked antibody (Cell Signaling, 7074) and Rhodamine (TRITC) Conjugated affinipure donkey antichicken (Jackson immunoresearch, 703-025-155) (1:500). Following washes with PBST, sections were incubated with TSA plus Fluorescein (1:250) (AKOYA biosciences, NEL741001KT). Afterwards, cellular nuclei were stained with DAPI (1:20000; Invitrogen D3571), washed again with PBST and finally mounted and coverslipped.

The sections were analyzed with an epifluorescence Leica DM6000 microscope (Leica, France) to check for the intrinsic fluorescence of the viruses and the identity of the infected cells. Mouse brains that didn’t meet the expression requirements led to the exclusion of the mice from the experiments. Micrographs were acquired at 10x (whole olfactory bulb) and 20x to later analyze protein co-expression.

Immunostaining For Electron Microscopy

WT and CB1-KO mice were deeply anesthetized by intraperitoneal injection of ketamine/xylazine (80/10 mg/kg body weight i.p.) and were transcardially perfused at room temperature (RT, 20-25 °C) with phosphate buffered saline (0.1 M PBS, pH 7.4) for 20 s, followed by a fixative containing 4% formaldehyde (freshly depolymerized from paraformaldehyde), 0.2% picric acid, and 0.1% glutaraldehyde in phosphate buffer (0.1 M PB, pH 7.4) for 10-15 min. Then, brains were removed from the skull and post-fixed in the same fixative for about 1 week at 4 °C. Afterwards, brains were stored at 4 °C in 1:10 diluted fixative solution until used.

Fifty µm-thick coronal OB sections were pre-incubated in blocking solution 10% bovine serum albumin (BSA), 0.1% sodium azide and 0.02% saponine prepared in Tris-hydrogen chloride buffered saline 1× (TBS), pH 7.4 for 30 min at RT. Then, sections from both WT and CB1-KO mice were incubated with a goat anti-CB1 receptor antibody (Frontier Institute Co., ltd; goat polyclonal; CB1-Go-Af450; FR100610, 1:100) and a guinea pig polyclonal anti-GLAST antibody (Frontier Institute Co., ltd; guinea pig polyclonal; GLAST-GP-Af1000; FR102170, 1:1,000) diluted in 10% BSA/TBS containing 0.1% sodium azide and 0.004% saponine on a shaker for 2 days at 4 °C. OB sections were incubated with 1.4 nm gold-conjugated rabbit anti-goat pig IgG antibody (Fab fragment, 1:100, #2006, Nanoprobes, Inc., Yaphank, NY, USA) and biotinylated donkey anti-guinea pig IgG antibody (1:200, 706-065-148, Jackson Immuno Research) diluted in 1% BSA/TBS 1x with 0.004% saponin on a shaker for 4 h at RT. Tissue was washed in 1% BSA/TBS 1x on a shaker at RT and incubated with ABC (1:50) prepared in washing solution for 1.5 h at RT. Sections were rinsed with 1% BSA/TBS 1x, stored overnight at 4 °C, and post-fixed with 1% glutaraldehyde in TBS 1x (1 ml/well) for 12 min at RT. After rinsing in double distilled water, gold particles were silver-intensified with the HQ Silver kit (#2012, Nanoprobes, Inc., Yaphank, NY, USA) in the dark for 12 min at RT. The OB sections were then washed with double distilled water and 0.1 M PB (pH 7.4) for 30 minutes. The biotinylated antibody was revealed with 0.05% DAB in 0.1 M PB (pH 7.4) containing 0.5% Triton X-100 and 0.01% hydrogen peroxide for 3.5 min at RT, followed by washes in 0.1 M PB (pH 7.4). Osmication was done with 1% osmium tetroxide in 0.1 M PB (pH 7.4) in the dark for 20 min. Sections were then washed, dehydrated in graded ethanol, cleared in propylene oxide, pre-embedded in a 1:1 mix of propylene oxide/Epon 812 resin overnight at RT, and finally embedded in pure Epon 812 resin. Electron micrographs were taken with a Hamamatsu FLASH digital camera inserted in a transmission electron microscope (JEOL JEM 1400 Plus).

Behavioral protocols

Social transmission of stress

Non-littermate animals were housed together at 3 weeks of age to establish a familiarity between them while avoiding dominance issues, and then moved to new cages in couples 1–2 days before the experiment. One of the members of the couple is the demonstrator (DEM, a C57BL/6-N mouse) while the other one is the observer (OBS, depending on the experiment: [i] wild-type CB1 flox mice, [ii] CB1 mutant mice, [iii] operated CB1-flox mice, [iv] operated C57BL/6-N) (Fig. 1a, b). As described previously⁸, the demonstrators were subjected to either a 5 min x 0.5 mA/30 s shock protocol (stress, foot-shock) in a clean fear conditioning chamber (stress DEMs), or to a 5-min separation in a novel cage (neutral) similar to the home-cage but with clean bedding (neutral DEMs), and they immediately moved back to the home-cage where they were allowed full interaction with the observer (Fig. 1a, b). Their behavior was recorded during 5 min and 8 different social and non-social behaviors were analyzed offline: anogenital exploration (snout toward the area of the congener), body exploration (all other snout contacts that are not on or near the anogenital region), allogrooming (grooming of the partner), self-grooming, digging, rearing, walking, sitting and fighting.

Odor-dependent social transmission of stress

To test whether odors were sufficient to induce transmission of stress, DEMs were habituated for three days to being swabbed on the anogenital region before the test with a clean cotton swab for 3 s, and a cotton swab was placed in the home-cage of each experimental couple for 2 days to avoid neophobia in the test. Demonstrators were swabbed with a humid cotton swab (wet with 1% saline solution) three times after the shock protocol (stress odor), or after being removed from the home-cage (neutral odor). A wet cotton swab was used as the control condition (wet swab) (Fig. 1m). Immediately after odor collection, the cotton swab was presented to the OBS in the home-cage, slightly touching their snout before dropping it on the cage bedding. Mice were allowed to interact with the cotton swab for 5 min, and then the cotton swab was removed from the cage. In the experiments of paired odor exposure and fiber photometry (Figs. 4, 5; Supplementary Fig. 4l, m, and Supplementary Fig. 6), the cotton swab was lightly maintained in front of the snout of the mouse during either 2 or 20 s before removing it from the cage.

Novel object recognition memory task

An L-shaped maze of gray PVC with two perpendicular arms placed on a white background was used in this test. The test was performed under at 50 ± 5lux intensity with an overhung camera allowing the recording and later offline scoring of the maze exploration by the mouse.

The test consists in 3 daily phases as described previously⁸⁷. On day 1, mice were habituated to the maze for 9 min before returning to the home-cage. On day 2, they were presented with two identical objects in each arm, and allowed to explore for 9 min to get familiar with them (acquisition phase). On day 3, mice were exposed to the maze again where one of the familiar objects is replaced by a novel one, and allowed exploration for 9 min. Exploration of an object was counted when the animal had the nose on the object or facing the object in a distance less than 0.5 cm. This phase tests the recognition performance of the animal by comparing the time spent in the novel versus the familiar objects. Object recognition capabilities are assessed by a discrimination index that is calculated by the time spent exploring the novel object minus the time spent exploring the familiar one, divided by the total exploration time. The position of the novel object and the associations of novel and familiar were randomized. All objects were previously tested to avoid biased preference. The apparatus as well as objects were cleaned with ethanol (70%) before experimental use and between each animal testing.

To test the STS effect on NOR acquisition, the couples of mice underwent the STS protocol 20 min before the acquisition phase on day 2 of the NOR test, and were tested the next day for NOR retrieval (Fig. 1i). To test the STS effect on NOR retrieval: animals underwent the STS protocol 20 min before the retrieval phase on day 3, and were subsequently tested for NOR retrieval (Fig. 1g). To study the long lasting consequences of STS, pairs of mice underwent STS on Day 3 (morning) and they were tested for NOR retrieval 6 h later (evening).

Social cognition test

Grouped house mice were habituated to an open field (30×30 cm) at 50 ± 5lux that contained two identical cages at opposite corners, with an object inside, for 3 min. After 10 min, one of the objects was replaced by an age-matched unfamiliar mouse from the same strain, and the tested animals were allowed to interact with the cages containing object and social stimuli for 3 min (social acquisition phase). After a 10 min ITI, in which the animals underwent the STS protocol, they were moved back to the arena were the object was replaced by a novel social stimulus (age-matched unfamiliar mouse), and allowed to explore for 3 min (social novelty phase). The social discrimination index was calculated as the time spent exploring the novel social stimulus minus the time spent exploring the familiar social stimulus divided by the total time of exploration.

Buried food test

As described previously for this test⁴⁰, mice were habituated to a food pellet for 3 days in the home-cage, and food deprived for 24 h up until the test. Animals were moved to home-cage sized cages with 3–5 cm of clean bedding and allowed to roam for 10 minutes for habituation. Then, they were removed from the cages momentarily and a food pellet was hidden below the bedding at a random corner. Mice were moved back to the cage and allowed to search for the pellet for maximum 5 minutes. The time of pellet retrieval was recorded offline and if this did not occur within 5 minutes, the test was considered as failed.

Odor detection test

The experiment was performed at 50 ± 5 lux in a 35.5 × 15 × 19 cm cage that had sawdust from the home cage of the animal, with a cover that had a hole to allow easier recording. The animals were habituated during 3 days prior the test. For the habituation, a mouse was place for 3 min in a cage that had a metal lid with a nozzle used to place a small piece of absorbent paper in it, in a way that will allow the odor impregnating the paper to diffuse, but not a direct interaction with the paper. During habituation, the piece of paper was impregnated with 10 µL of sesame oil, and the mice were subjected to 5 trials of 3 min each with 3 min inter trial, in which the mice were returned to their home cage and the paper was replaced with a similar one.

The fourth day, the test was performed. The animals were food deprived 24 hours before the test. In the test, two odors, benzaldehyde (almond) and isoamylacetate (banana) (both from Sigma Aldrich, France), were chosen to analyze the capability of detection between different odor concentrations. In the first trial 10 µL of sesame oil were placed under the metal cover. In the subsequent trials, 10 µL of benzaldehyde (almond) or isoamylacetate (banana) at the concentrations 0.001%, 0.1%, 0.1% and 1% were placed under the cover, in trials 2, 3, 4 and 5, respectively. The metal cover was cleaned thoroughly between trials with ethanol 30%, so as the testing cage. The time spent sniffing the central part of the metal cover was counted offline, and counted as odor exploration while the time spent sniffing the metal cover by itself was not counted as it was assumed to be object exploration.

Odor discrimination test

The experiment was performed at 50 ± 5 lux in a new cage identical to the home-cage with clean sawdust, covered by a plexiglass with holes that allowed the exposure to a cotton swab hanging over the cage. The animals were habituated to the cage for 10 min with a wet swab. Then, the cover was quickly removed to replace the swab with a novel one with 40 µL of benzaldehyde 0.05% (almond), and animals were allowed to explore the swab for 2 min. This was repeated two more times with new swabs, with an intertrial time of 1 min. After that, novel swabs with 40 µL of isoamylacetate 0.05% (banana) were presented for 3 consecutive times. The time spent sniffing the swabs during the 3 minutes was counted as odor exploration.

Elevated plus maze test

The test was performed in an elevated plus maze consisting of 4 arms (height: 66 cm) of 45-cm long and 10-cm wide disposed cross-shaped and connected by a central platform of 10 cm × 10 cm. The open arms had a light intensity of 75lux and the closed arms of 20 lx. OBS mice were placed in the open platform 20 min after social interaction with demonstrators, and allowed to explore the maze for 5 min. The time spent in open and closed arms, and the number of times they enter in those, was analyzed offline by an experimenter blind to the condition.

Social interaction with a stranger test

Experimental animals were habituated in their home-cage for 10 min to the testing room. Then, a C57BL/6-N mouse of the same sex and age was introduced in the home-cage for 5 minutes, allowing full interaction between resident and stranger. Videos of the social interaction were recorded and 8 behaviors of the resident animals towards the partners were analyzed offline: anogenital exploration, body exploration, allogroom, self-groom, digging, rearing, walking, sitting and fighting. Animals that exhibited aggressive behaviors for more than 1 min were excluded.

Fiber Photometry

Five to six weeks after surgery, freely-moving mitoGcAMP6s/cytosolic GAMP6f-expressing mice were imaged using 470 nM LED to excite the sensor, and 405 nM for the isosbestic signal control. Observer mice with fiber implants were habituated to the connection during 3 days prior the test, in 10-min sessions in which they were connected and allowed to roam in the home-cage with their familiar cage-mate. The fiber photometry set-up collected the emitted fluorescence with a sCMOS camera (Hamamatsu Orca Flash v3) through an optic fiber (core 400 μm, N.A 0.5) divided in 2 sections: a short fiber implanted in the brain of the mouse and a long fiber (modified patchcord), both connected through a ferrule-ferrule (1.25 mm) connection. To minimize the photobleaching effect of the recording and preserve a high signal to noise ratio, the light intensities in the tip of the patch cord were adjusted to ∼100 μW for the 470 nm channel and ∼50 μW for the 405 nm channel. A custom MATLAB script (Matlabworks) was used to synchronize video recording with fiber photometry, combined with a programmed Arduino board. The sampling rate was settled at 20 Hz for both photometry (interleaved) and video recording.

On the test day, observer mice were separated from their partners for a habituation period of 5 min in which they stayed in the home-cage while their partners were shocked (stress odor) or just separated (neutral odor) to collect the odor in a cotton swab. Observers were exposed to a cotton swab wet or the cotton swab impregnated with the partner chemosignals in an inter-individual alternated order (some mice had saline first, others the social odor first) with an interval of 4–5 min. Mice were only exposed to each odor one time, by establishing a close contact between the impregnated swab and the snout during an average time of 20 s before removing the swab from the cage. Ca²⁺ signals were recorded during the duration of the test (20 min). For non-social neutral odors odors, the animals were presented with a wet swab, a swab containing 40uL of isoamycetate 0.05% (Sigma) and a swab containing benzaldehyde 0.05% (Sigma) in an inter-individual alternated order with an interval of 4–5 min, while mitochondrial calcium changes were recorded.

Raw calcium Ca²⁺ were pre-processed by removing the first minute of the recording to decrease the effect of the first exponential photobleaching, and by removing point artifacts. The 470 nM signal was fitted to the isosbestic 405 nM using a polynomial fit of first degree and, for each time point, ΔF/F was calculated as (F470nm – F405nm(fitted))/F405nm(fitted). ΔF/F values were smoothed using a moving average of 0.5 s. Z-score was calculated in the whole recording to take into account the changes in signal intensity during the experiments. The signal corresponding to 1 min after the onset of the odor exposure was extracted, with 15 s baseline before the onset. The baseline values were used to correct the extracted signal by performing a subtraction of the mean of the baseline to the whole extracted signal. The area under the curve during the 20 s of swab exposure of the odors was calculated from the z-scored data. Ca²⁺ signals from wet swab and the different odors were compared within mice.

Quantification and statistical analysis

Data collection

All data points that appear in the graphs of this study correspond to individual sample mice, and not technical replicates. Statistical methods to determine sample size were not used, but the numbers of animals used were similar to those in the literature. Experimenters analyzing the raw videos were always blind to the conditions of the subject. All mice were randomly assigned to experimental conditions. We used custom software to analyze the social behaviors and time spent in each arm during the novel object recognition. For the analysis of the immunostaining, we used FIJI. The contrast and brightness parameters were adjusted and applied equally to all micrographs. FIJI’s cell counter plug-in was used to establish overlap between differently expressed proteins. For the fiber photometry data, we extracted the signal as described above using the provided custom code. Raw data from all experiments was processed and analyzed using Microsoft Excel 2020 and Graph Pad 8.0.

Statistical analysis

Graphs and statistical analysis were performed with Graph Pad 8.0. All data come from distinct samples (individual mice) and they are shown as independent data points per animal ± standard error of the mean (SEM). Each experiment was repeated with at least two independent batches. Normality of the data was assessed with the Kolmogorov-Smirnoff test for all sample sizes >5 or Shapiro-Wilk test for sample sizes t test, ordinary one-way ANOVA with Bonferroni post hoc analysis, or ordinary two-way ANOVA when necessary) or non-parametric (unpaired Mann-Whitney test, Wilcoxon matched pairs signed rank test or Kruskal-Wallis test with Dunn’s post hoc analysis) were performed. Detailed statistical data for each experiment including exact mean ± SEM values, test statistic with confidence intervals, group sizes, degrees of freedom and exact P values can be found in Supplementary Table 1 (for main Figures) and Supplementary Table 2 (for Supplementary Figs.).

Reporting summary

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