Experiments were performed on adult (3–6 months old) C57BL/6J mice (Janvier Laboratories, France). Animals were housed in animal care facilities (PREBIOS, University of Poitiers; UEPAO, Centre INRAe Val-de-Loire). Experimental procedures were carried out in accordance with the guidelines of the French Ministry of Agriculture and the European Communities Council Directive and validated by the Regional Ethical Committee (Authorization no. 02184.01).

SKF-96365 and YM-58483 were purchased from Sigma Aldrich (Saint-Louis, MO, USA) and dissolved in water and DMSO, respectively. GSK-7975A from Aobious (Gloucester, MA, USA) was dissolved in DMSO. Leptin was purchased from R&D Systems (Minneapolis, MN, USA) and dissolved in water. Cell culture media and growth factors were from Invitrogen (Carlsbad, CA, USA).

Area postrema was microdissected out of C57BL/6J adult mice brains, incubated in accutase for 15 min at 37°C and then mechanically dissociated as single cells that were grown in DMEM-F12 supplemented with 1% B27 and with 20 ng.ml^–1 EGF and 10 ng.ml^–1 FGF2. Neurospheres were grown in suspension and obtained after 1 week in vitro (Figure 1A). For replating, neurospheres were collected, mechanically dissociated as single cells and seeded to obtain secondary neurospheres that were used for Ca²⁺ imaging, RT-PCR, western blotting, immunostaining and cell proliferation assays.

For western blotting analysis, area postrema neurospheres were lysed in Laemmli loading buffer (Sigma Aldrich) and heat denatured at 95°C for 5 min, in parallel with subventricular zone neurospheres used as positive controls (Domenichini et al., 2018). All samples were resolved in a 9% sodium dodecyl sulfate-polyacrylamide gel along with pre-stained standard molecular weight markers (BioRad Laboratories, Hercules, CA, USA). Proteins were transferred to nitrocellulose membranes (0.20 μm-pore size; GE Healthcare, Little Chalfont, UK), using a mini protean electroblotter (BioRad Laboratories). After three washes in TBS (20 mmol.l^–1 Tris–HCl, 150 mmol.l^–1 NaCl, pH 7.5) containing 0.1% (volume/volume) Tween 20 (TBS-Tween), immunoblots were probed overnight at 4°C in TBS-Tween with 3% (weight/volume) fat milk with either rabbit polyclonal anti-Orai1 (H-46, Santa Cruz Biotechnology, Dallas, TX, USA; 0.4 μg.ml^–1), or mouse monoclonal anti-TRPC1 (E-6, Santa Cruz Biotechnology; 0.4 μg.ml^–1), or mouse monoclonal anti-STIM1 (GOK 610954, BD Biosciences, Franklin Lakes, NJ, USA; 0.4 μg.ml^–1), or rabbit polyclonal anti-actin (A2066, Sigma Aldrich; 1/5000). Membranes were then incubated for 1 h at 4°C with anti-mouse or anti-rabbit immunoglobulin-HRP-linked (NA931V and NA934V, respectively, 1/5000; GE Healthcare) before detection of the bound antibodies using the ECL enhanced chemiluminescence system (Immobilon, Millipore, Billerica, MA, USA). Results were analyzed with GeneGnome XRQ (SYNGENE Ozyme, Cambridge, UK).

Immunostaining was performed on the one hand, on area postrema neurospheres thrown on glass slides with a cytospin and fixed in methanol at −20°C, and on the other hand, on adult mouse brain sections. For brain sections, mice were deeply anaesthetized and transcardially perfused with 0.9% NaCl solution followed by phosphate-buffered 4% paraformaldehyde solution at 4°C during 10 min. Brains were removed, post-fixed for 1 h at 4°C, and cryoprotected in phosphate-buffered 30% sucrose solution at 4°C overnight before cutting 20 μm-thick sections using a cryostat (Leica 2500).

For immunostaining, preparations were permeabilized and non-specific binding sites were blocked prior to incubation with a goat polyclonal anti-SOX2 (Y-17, Santa Cruz Biotechnology; 2 μg.ml^–1) or a mouse monoclonal anti-SOX2 (A-5, Santa Cruz Biotechnology; 2 μg.ml^–1) antibody along with one of the following antibodies: rabbit polyclonal anti-Orai1 (H-46, Santa Cruz Biotechnology; 2 μg.ml^–1), or mouse monoclonal anti-TRPC1 (E-6, Santa Cruz Biotechnology; 2 μg.ml^–1), or mouse monoclonal anti-STIM1 (GOK 610954, BD Biosciences; 2 μg.ml^–1). Preparations were then revealed with the appropriate Alexa fluor 555 or Alexa fluor 488 conjugated antibodies (1/200, A21432, A31570, A21206, A21202; Invitrogen). DAPI was used to label nuclei. Preparations were then analyzed with an FV-1000 spectral confocal station installed on an inverted microscope IX-81 (Olympus). The emitted fluorescence was detected by spectral detection channels between 425–475 nm, 500–530 nm, and 550–625 nm, for UV, green and red fluorescence, respectively.

Neurospheres were dissociated and 30,000 cells were plated as single cells at on fibronectin-coated glass coverslips. After 2 h, the cells were incubated for 30 min at 37°C with 3 μmol.l^–1 of the Ca²⁺ sensitive probe Fura-2AM (Santa Cruz Biotechnology). Assays were performed in standard external buffer solution (130 mmol.l^–1 NaCl, 5.4 mmol.l^–1 KCl, 0.8 mmol.l^–1 MgCl₂, 10 mmol.l^–1 HEPES, 5.6 mmol.l^–1 D-glucose, pH 7.4) supplemented with 1.8 mmol.l^–1 Ca²⁺ or 0.1 mmol.l^–1 EGTA for the Ca²⁺- free solution. Cells were perfused with the buffer solution containing 0.1 mmol.l^–1 EGTA (Ca²⁺-free solution) and SOCE was triggered with thapsigargin (Sigma Aldrich, 4 μmol.l^–1). For experiments with leptin, area postrema cells were incubated with 6.25 nmol.l^–1 leptin during 30 min at 37°C prior to intracellular Ca²⁺ measurements. Fluorescence images were recorded using a lambda 421 coupled to an Olympus IX73 inverted microscope and a Zyla sCMOS 4.2 PLUS camera. Consecutive excitation of Fura-2AM at 340 nm and 380 nm was undertaken every 2 s, and emission fluorescence was collected at 505 nm at 37°C using the Metafluor software. The 340/380 nm fluorescence ratio was measured over time in selected regions of interest (ROI) on several cells whose numbers are given in Figures 2B, C, 4B. The fluorescence ratio (340/380 nm) was measured in each cell and subtracted from the value obtained after store depletion, just at the time of Ca²⁺ re-addition (This value is expressed as Δ ratio 340/380 nm on the figures).

Cell proliferation was determined by a bromodeoxyuridine (BrdU) incorporation assay. To this aim, area postrema neurospheres were dissociated as single cells and plated at the concentration of 30,000 cells.ml^–1 in poly-L-Lysine-coated 96 well plates. The cells were treated for 24 h with SKF-96365, YM-58483 or GSK-7975A used at the concentrations ranging, respectively, from 10 nmol.l^–1 to 10 μmol.l^–1, 100 nmol.l^–1 to 10 μmol.l^–1 or 1 μmol.l^–1 to 30 μmol.l^–1, as indicated in Figure 3A. BrdU was added to the medium for the last 4 h of the assay. BrdU incorporated into DNA was quantified according to the manufacturer’s instructions (Roche Diagnostics, Meylan, France).

The area postrema was dissected from mouse brains, dissociated, and seeded at 10,000 cells per ml in 24-well plates in cell culture medium supplemented with any of the following treatments: SKF-96365 (1 μmol.l^–1), YM-58483 (5 μmol.l^–1), GSK-7975A (20 μmol.l^–1) or leptin (6.25 nmol.l^–1). Controls were performed with culture medium alone (for SKF-96365 and for leptin) or supplemented with DMSO (solvent of YM-58483 and GSK-7975A that was diluted like the drugs i.e., 1/10,000 for YM-58483 and 1/2,500 for GSK-7975A). Primary neurospheres were allowed to grow during 1 week and then counted under a microscope. For self-renewal assays, area postrema primary neurospheres were collected, dissociated as single cells, and reseeded in culture medium to obtain secondary neurospheres.

Statistical significance of differences was examined by one-way analysis of variance (ANOVA) followed by the Bonferroni’s post-hoc test for multiple comparisons, or by non-parametric Mann and Whitney test for pairwise comparisons (Statview 5.00 software). The statistical significance level was set for p-values < 0.05 and represented on the figures by: *p < 0.05, **p < 0.01, ***p < 0.001.

Mainly consisting of Orai1, SOCs are activated by the Ca²⁺ sensor STIM1 at the ER membrane and can mobilize also TRPC1 which will amplify the Ca²⁺ entry. Western blot (Figure 1A’) analysis shows that neurospheres from area postrema express TRPC1, Orai1 and STIM1 (Figure 1A). These results were confirmed by immunostainings (Figures 1B–D), which furthermore highlights the presence of SOC molecular actors in neural stem-like cells expressing SOX2. The physiological relevance of these data obtained in vitro was then investigated on mice brain sections. Figures 1E–G and their corresponding magnifications in Figures 1E’–G’ illustrate that TRPC1, Orai1 and STIM1 are found in area postrema cells in vivo within SOX2-expressing stem cells.

Next, we assessed the functionality of SOCs in area postrema NSCs by determining their ability to support Ca²⁺ entries in response to depletion of ER Ca²⁺ stores. To this end, cells were loaded with Fura-2AM, a Ca²⁺ sensitive ratiometric dye that binds to free intracellular Ca²⁺. To deplete ER Ca²⁺ stores, we used thapsigargin that blocks the cell’s ability to pump Ca²⁺ into the ER through the sarco-endoplasmic reticulum calcium-ATPase (SERCA). Figure 2A (first peak) shows that exposure to thapsigargin (TG) in a zero Ca²⁺ solution induces an increase in intracellular Ca²⁺ concentration in area postrema NSCs due to inhibition of Ca²⁺ uptake and a passive leak of Ca²⁺ from the ER. The ER store-depletion is sensed by STIM1 and secondarily activates SOCs at the plasma membrane. Although SOCs open within a few tenths of a second after Ca²⁺ release from the endoplasmic reticulum, no Ca²⁺ can enter the cell as long as the cell is maintained in a Ca²⁺-free buffer solution. After Ca²⁺ release from ER was over, Ca²⁺ was added to the extracellular medium and permitted a Ca²⁺ influx into the cell (SOCE, Figure 2A, second peak). To quantify the SOCEs, the initial slopes observed just after addition of Ca²⁺ were measured on traces obtained from multiple cells. In addition, we quantified the maximum value of the second peak, even though this parameter depends both on Ca²⁺ entries through SOCs, on processes that extrude Ca²⁺ from cells and on uptake in the mitochondria. Quantification of the slope and maximum of response showed a decrease of about 35% for GSK-7975A (5 μmol.l^–1) and 46% for YM-58483 (also known as BTP2; 1 μmol.l^–1) compared to control (DMSO). These results indicate that SOC inhibitors decreased SOCE, confirming the presence of functional SOCs in area postrema cells (Figures 2B, C).

Taken together, our data establish that area postrema NSCs are endowed with functional SOCs.

As SOCs are known to influence neurogenesis in the subventricular zone, we investigated whether they could regulate NSC proliferation and self-renewal in area postrema by determining the impact of their pharmacological blockade. Cell proliferation was assessed by measuring BrdU incorporation into DNA during the S phase of the cell cycle. For this assay, the SOC inhibitors, YM-58483, GSK-7975A, and SKF-96365 were used (Li et al., 2012; Aulestia et al., 2018; Domenichini et al., 2018; Terrié et al., 2021). The results shown on Figure 3A demonstrate that SKF-96365, YM-58483 and GSK-7975A, dose-dependently decrease area postrema cell proliferation for concentrations ranging from 0.1 to 10 μmol.l^–1, and 0.1 to 10 μmol.l^–1, and 3 to 30 μmol.l^–1, respectively, which is within the range of concentrations required for these inhibitors to induce cellular effects (Prakriya and Lewis, 2015).

Neural stem cells are characterized in vitro by their capacity to form neurospheres and self-renew when maintained in the presence of mitogens. To investigate the possible regulatory role of SOCs on the activation and self-renewal of area postrema NSCs, freshly dissociated area postrema cells were plated at the concentration of 10,000 cells per ml in the absence (control) or presence of the SOC inhibitors SKF-96365, YM-58483, and GSK-7975A (Figure 3B) used at concentrations of 1, 5, and 20 μmol.l^–1, respectively, which still allow mitotic activity (Figure 3A). Neurospheres were allowed to develop for 1 week and then counted. The number of neurospheres (primary neurospheres indicated as SPHERE I in Figure 3B) obtained reflects the activation of NSCs in the culture. Primary spheres encompass progenitors (represented in white on Figure 3B) resulting from stem cell proliferation, and stem cells (depicted in black in Figure 3B) resulting from self-renewal, i.e., proliferation and maintenance of a stem cell phenotype. To quantify stem cells in primary neurospheres and thus analyze the self-renewal process, we determined the number of new neurospheres at the next generation (Azari et al., 2011). To this aim, primary neurospheres obtained in each experimental condition were, respectively, collected, dissociated, plated and allowed to develop into secondary neurospheres (Figure 3B, SPHERE II) in the culture medium without addition of SOC inhibitors. Our results show that the number of primary neurospheres (Figure 3C) and secondary neurospheres (Figure 3D) are dramatically reduced by the addition of SKF-96365, YM-58483, and GSK-7975A, indicating that SOCs play a major role in both mitotic stimulation and self-renewal of NSCs in the area postrema.

Store-operated Ca²⁺ entries are evoked and modulated by various extracellular signals, including hormones (Dragoni et al., 2011; Rodríguez-Moyano et al., 2013; Somasundaram et al., 2014; Zuccolo et al., 2018; Núñez et al., 2019; Lv et al., 2020). Among them, leptin is an adipose tissue-derived hormone that controls food intake and energy expenditure by acting on peripheral organs and brain structures (Morton et al., 2014). The area postrema plays a major role in leptin control of energy homeostasis (Hayes et al., 2010). Leptin used at physiological nanomolar concentrations has been shown to reduce Ca²⁺ influx in several cell types, including neurons (Seufert et al., 1999; Fortuño et al., 2002; Shanley et al., 2002; Kohno et al., 2007, 2008; Sun et al., 2019).

As we observed that area postrema NSCs express the leptin receptor (Supplementary Figure 1), a possible effect of leptin on SOCs was investigated. SOCEs were measured in cells preincubated with 6.25 nmol.l^–1 leptin during 30 min before Ca²⁺ imaging. Figure 4A depicts a Ca²⁺ trace obtained in the absence (full line) or presence (dashed line) of leptin. Quantification of the initial slope and maximum of the traces reveals that leptin significantly reduces SOCEs of 54 and 15%, respectively, compared to control (Figure 4B).

Consistent with the effects of leptin on SOCEs (Figures 4A, B) and the data of Figures 3C, D showing that an inhibition of SOCs results in decreased activation of area postrema NSCs, leptin was found to reduce both sphere formation and self-renewal of area postrema NSCs (Figure 4C). Furthermore, co-treatment of area postrema NSCs with leptin and the SOC inhibitor YM-58483 amplified neither YM-58483 effects SOCEs (that was decreased by about 50% for the slope in both YM-58483 and YM-54843 + leptin and between 70 and 50% for the maximum of response for YM-58483 and YM-54843 + leptin, respectively, compared to control) (Figure 4D) nor its effects on neurospheres (Figure 4E), suggesting leptin exerts its effects on area postrema NSCs by inhibiting SOCs. Regarding the maximal response, it was unexpectedly slightly higher when YM-58483 was combined with leptin than when YM-58483 was used alone, which may be related to the fact that the maximal response depends on Ca²⁺ entries through SOCs and could secondarily modulate other Ca²⁺ processes.