C57BL/6 J male mice (7 weeks old) were used for experiments in which adeno-associated viruses (AAVs) expressing shRNAs designed to inactivate P2X7Rs in a cell type-specific manner were injected intracerebroventricularly (i.c.v.). The respective controls were injected via the same route with AAVs expressing inactive shRNAs (see below). The astrocyte-specific deletion of P2X7Rs in mice was also achieved by breeding conditional humanized P2X7R mice P2rx7tm1.1(P2RX7)Jde (MGI:6203041; P2rx7flox/flox; [16]) to the astrocyte-specific Cre driver line B6.Cg-Tg(GFAP-Cre/ERT2)505Fmv/J (Strain #012849; The Jackson Laboratory, Bar Harbor, ME, USA). We crossed homozygous humanized mice bearing floxed P2rx7 alleles to GFAP-Cre/ERT2 mice to obtain P2rx7flox/+; GFAP-Cre/ERT2 mice, which were bred to P2rx7flox/flox mice to generate P2rx7flox/flox; GFAP-Cre/ERT2 mice. The subsequent experiments were performed by using offsprings of P2rx7flox/flox mice crossed with P2rx7flox/flox; GFAP-Cre/ERT2 mice. To achieve astrocyte-specific P2X7R inactivation, mice received intraperitoneally (i.p.) 50 µg/g body weight tamoxifen (Sigma-Aldrich, St. Louis, MO, USA) dissolved in corn oil, every 12 h for five consecutive days. Behavioral experiments were conducted 2 weeks later. The animal study was reviewed and approved by the Institutional Review Board of the Chengdu University of TCM, Chengdu, China (protocol code DC2742, 14 February 2022).

To knock down murine P2X7R (mP2RX7) expression in specific cell types, different adeno-associated viruses (AAVs) were applied expressing shRNA directed against mP2RX7 (5′-GCGGAAAGAGCCTGTTATCAG-3′) or control shRNA which are driven by cell type-specific promoters. The following viruses were purchased from Guangzhou PackGene Biotechnology Co., Ltd. (Guangzhou, China): neuron-specific (driven by human synapsin promoter): AAV9_hSyn-EGFP-miR30shRNA(mP2X7)-WPRE-SV40pA and AAV9_hSyn-EGFP-miR30shRNA(NC)-WPRE-SV40pA; astrocyte-specific (driven by 681 bp GFAP promoter): AAV8_GFAP(681 bp)-EGFP-miR30shRNA(mP2X7)-WPRE-SV40pA and AAV8_GFAP(681 bp)-EGFP-miR30shRNA(NC)-WPRE-SV40pA; and microglia-specific (driven by Iba1 promoter): AAV9_Iba1-EGFP-miR30shRNA(mP2X7)-WPRE-SV40pA and AAV9_Iba1-EGFP-miR30shRNA(NC)-WPRE-SV40pA. The oligodendrocyte-specific shRNA (driven b32.y Olig1 promoter): AAV2_Olig001-H1-shRNA(mP2RX7)-CAG-EGFP-WPRE-pA and AAV2_Olig001-H1-shRNA(Luciferase)-CAG-EGFP-WPRE-pA were purchased from Shanghai Teleng Biotechnology Co., Ltd. (Shanghai, China).

Mice were anesthetized with isoflurane (5% induction; 2% maintenance; RWD Life Science, San Diego, CA, USA) and fixed on a stereotactic platform (RWD Life Science). A 5 µl microsyringe was inserted into the right lateral ventricle at the following stereotactic coordinates (− 0.4 mm anterior–posterior, 1.0 mm lateral to the bregma, 3.0 mm below the surface of the skull) [17]. Cell-specific shRNAs were injected in a volume of 3 µl and at a rate of 0.1 µl/min. The original virus titer of all shRNAs was greater than or equal to 1E + 13 VG/ml; before i.c.v. application, they were diluted with PBS 2.5 times. At the end of surgery, 5 mg/kg enrofloxacin (Bayer, Leverkusen, Germany) was administered subcutaneously (s.c.) to the animals to prevent postoperative infection. An interval of 2 weeks was kept between shRNA application and the start of behavioral or immunohistochemical experiments.

TST, FST, and SPT, the latter after inescapable foot shock, were conducted as described earlier (see present Fig. 1 [18]). In short, the first two of these tests represent an acute stress for the animals and induce after a certain time-lag, behavioral despair. Thus, after being suspended by their tails for 6 min, mice try to escape from this position, but after a few minutes they give up their futile attempts and remain temporarily immobile. The cumulative immobility time is determined as a measure of the depressive-like behavior. A similar reaction evolves in the FST, in which mice swim in water, filled into a glass cylinder. The duration of cumulative immobility within the 6-min observation time is also determined in this case.

Depressive-like reactions determined in the tail suspension (TST) and forced swim tests (FST) in mice depend on the presence of astrocytic P2X7Rs; anhedonia induced by foot shock followed by the measurement of sucrose consumption is not changed in the absence of P2X7Rs in the four major cell-types of the CNS. A TST, FST, and sucrose consumption following foot shock stimulation were carried out in this order (Group I) or the reverse order (Group II). The sequence of measurements did not modify the immobility time or sucrose preference. B P2X7R deficiency in astrocytes resulted in a marked shortening of the immobility time (TST, FST) but did not affect sucrose consumption in the SPT. C The i.c.v. injection of shRNA specific for astrocytic P2X7Rs had a similar, although less pronounced effect on the TST/FST and no effect on SPT. Injection of shRNA selectively inactivating P2X7Rs in neurons (D), oligodendrocytes (E), and microglia (F) did not change any of the depressive-like reactions. Each column designates the mean ± SEM of 10 mice and in addition shows the individual values by separate dots. *P < 0.05; statistically significant difference from the wild-type mice (B) or from the respective control shRNA-treated group (C; paired t-test or Mann–Whitney rank sum test, as adequate)

We applied foot shock with a current intensity of 1 mA in a “Learned Helplessness Test system” (Shanghai Xin-Ruan Information Technology Co., Shanghai, China); each stimulus lasted for 5–15 s with a random interval in between. On the same day, two 50-ml bottles were given to the mice, one containing 1% sucrose solution and the other one containing water. On the 2nd and 3rd days again sucrose solution and water were given, but each day by changing the position of the bottles to avoid a side bias. On the 4th day, water and food were deprived. During the night of the 5th day, sucrose preference was determined in % as (volume sucrose/(volume sucrose + volume water)) × 100%. After foot shock stimulation, the mice became “anhedonic” with a decrease in the consumption of the sweet solution.

Mice were sacrificed with the i.p. injection of 1% sodium pentobarbital (0.4 ml) and were subsequently transcardially perfused with normal saline followed by pre-cooled paraformaldehyde (4% w/v in phosphate-buffered saline (PBS)). The brains were removed and postfixed in 4% paraformaldehyde overnight. Then, they were dehydrated with gradient (20–30%) sucrose in PBS at 4 °C. Coronal 40 μm-thick sections were prepared and incubated in a cryostat (Leica CM1860, Leica Biosystem, Muttenz, Switzerland) at − 20 °C until use. The sections were afterwards incubated in a blocking solution containing 4% bovine serum albumin (Sigma-Aldrich) and 0.5% Triton X-100 (Solarbio, Beijing, China) for 2 h at room temperature. Subsequently, the sections were incubated with mouse anti-GFAP (1:500; Cell Signaling Technology, Boston, MA, USA) and mouse anti-NeuN (1:1000; Abcam, Cambridge, MA, USA) overnight at 4 °C in the blocking solution. After washing with 0.1 M PBS containing 0.1% Triton X-100, the sections were incubated for 2 h with goat anti-mouse AF594 secondary antibodies (1:200, Bioss, Beijing, China) and DAPI (1:10,000, Abmole Bioscience, TX, USA) in the blocking solution at room temperature, washed three times with 0.1 M PBS containing 0.1% Triton X-100, and then mounted on slides using coverslips.

Image acquisition was performed using a confocal laser scanning microscope (Olympus IXplore SpinSR, Olympus, Tokyo, Japan). The number of GFAP/NeuN immunopositive cells and glial fibrillary acidic protein (GFAP)/NeuN immunopositive cells co-localized with EGFP, was manually determined by using the Cell Counter function of the ImageJ software within a defined region of interest (ROI) in each tissue section (see Fig. 2C). The percentage of EGFP+-GFAP+ or NeuN+-GFAP+ cells, in relation to the total GFAP+ and NeuN+ cell numbers, respectively, was calculated in the given ROI.

Immunohistochemistry and autofluorescence indicate in the mouse hippocampal CA1 region the co-expression of the astrocytic marker glial fibrillary acidic protein (GFAP), but not the neuronal marker NeuN, with EGFP contained in the vector of the astrocyte-specific P2X7R-shRNA. A DAPI fluorescence in the cell nuclei, GFAP immunopositivity in the cell bodies, and EGFP autofluorescence in the cell bodies, are co-localized as documented by their overlay in the merged pictures. B Analogous measurements show no co-staining for NeuN immunopositivity and EGFP autofluorescence. C Statistical evaluation of four similar experiments for each condition. The percentage of EGFP+-GFAP+ or NeuN+-GFAP+ cells, in relation to the total GFAP+ and NeuN+ cell numbers, respectively, was calculated in the given region of interest (ROI; within the limits of the broken lines). Each column designates the mean ± SEM of four experiments with individual values illustrated as separate dots. *P < 0.05; statistically significant difference between the two columns (Mann–Whitney rank sum test). Two different magnifications in the upper and lower rows of panels A, B

All data were expressed as means ± SEM of n observations, where n means the number of animals both in the behavioral and immunofluorescence studies. GraphPad Prism10 was used for the construction of figures and for statistical evaluations. We tested whether the sampled distribution of data satisfied the normality and equal variance criteria. Comparison of two data sets was made by the parametric Student’s t-test or the non-parametric Mann–Whitney rank sum test, as appropriate. A probability level of 0.05 or less was considered to be statistically significant.