The ADORA2A-knockout (gKO) mice used in the study were obtained from Dr. Jiang-Fan Chen (Molecular Neuropharmacology Laboratory, Boston University School of Medicine, Boston, MA, USA). Wild-type (WT) C57 mice were purchased from the Department of Laboratory Animal Science, Army Medical University, China. All mice in this study were 8–9 weeks old and weighed 25–28 g. All animal experiments and care were approved by the Third Military Medical University Committee on Ethics in the Care and Use of Laboratory Animals.

The ADORA2A-specific agonist CGS21680 (Cat. No. 1063) and -specific inhibitor SCH58261 (Cat. No. 2270) were purchased from Tocris Bioscience (Abingdon, UK). The STAT3-specific activator Colivelin (sc-361153) and -specific inhibitor Stattic (sc-202818) were purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX, USA).

A mouse model of chronic cerebral hypoperfusion (CCH) was established by referring to previously described methods (7). After placing the mouse under isoflurane inhalation anesthesia, the bilateral common carotid arteries were fully freed. Then, two 3-0 nylon sutures were loosely looped around the proximal and distal ends of the right common carotid artery, which was subsequently suspended using hemostatic forceps and placed in a special mini coil with an inner diameter of 1.8 mm. The coil was then carefully placed around the common carotid artery. Another coil was placed around the left common carotid artery in the same way 30 min later. The mice in the sham group were subjected to the same surgery except that the bilateral common carotid arteries were not surrounded with the coils. During the operation, the rectal temperature of the mouse was maintained at 36.5–37.5°C. Additionally, the 418-1 probe of a laser Doppler flow meter was fixed at the junction of the temporal squama and the greater wing of the sphenoid above the zygomatic arch to measure the cerebral blood flow and to evaluate the modeling throughout the entire surgical process.

The following five groups were created for the first part of the animal experiments: (1) sham group, (2) CCH group, (3) CCH+SCH58261 group, (4) CCH+CGS21680 group, and (5) CCH+CGS21680 (KO) group. For the first four groups, WT mice were modeled as described above and then randomly assigned to groups 1, 2, 3, and 4. The mice in the SCH58261 group were intraperitoneally injected with 0.1 mg/kg SCH58261, those in the CGS21680 group were intraperitoneally injected with 0.25 mg/kg CGS21680, and those in the CCH and sham groups were intraperitoneally injected with the same volume of dimethyl sulfoxide (DMSO) (14), respectively. For the fifth group, gKO mice were also modeled as described above and then intraperitoneally injected with 0.25 mg/kg CGS21680.

In the second part of the animal experiment, WT mice were modeled and divided into the following eight groups: (1) sham group, (2) CCH group, (3) CCH+Stattic group, (4) CCH+Colivelin group, (5) CCH+CGS21680 group, (6) CCH+CGS21680+Colivelin group, (7) CCH+ SCH58261 group, and (8) CCH+SCH58261+Stattic group. Stattic was intraperitoneally injected into the designated groups at a concentration of 2.5 mg/kg (15) and Colivelin likewise at a concentration of 1 mg/kg (16). All drugs were administered once every 24 h until the sudden death of the animals.

An the end of 2nd, 4th, and 6th weeks after CCH, five mice were taken from each group, deeply anesthetized with 1% sodium pentobarbital (60 mg/kg), and then perfused with 4% paraformaldehyde. Subsequently, the brain tissue was excised, fixed, dehydrated, and frozen. Then, the frozen brain sectioned (10-μm-thick slices) were incubated with 3% H₂O₂ for 10 min and blocked with goat serum for 30 min. Then, a primary antibody solution against myelin basic protein (MBP) (1:200 dilution; Abcam, Cambridge, MA, USA) was added for overnight incubation at 4°C, following which a sheep anti-rabbit IgG secondary antibody solution (1:200; Zhongshan Biotechnology, Beijing, China) was added for 30 min incubation at ambient temperature. Finally, nickel ammonium sulfate was used to enhance the color development, and Image Pro Plus software was used for semi-quantitative analysis of the staining results. A semi-quantitative analysis was used to assess the MBP⁺ cells as previously reported (14). Briefly, the MBP⁺ cells in a 0.5 mm × 0.5 mm width and 0.25 mm²/field of view were counted. The MBP⁺ cells counterstained with DAPI in nuclear were included for analysis.

Primary astrocytes were cultured according to previously described methods (17). The cerebral hemisphere of mice born within 3 days was excised and the pia mater was stripped off. The tissue was minced and then digested with 0.125% trypsin for 8 min, following which F12 complete medium containing 10% fetal bovine serum was added to stop the digestion. After centrifugation, the cell pellet was resuspended in F12 complete medium containing 100 μg/mL penicillin and 100 μg/mL streptomycin and transferred to a 75 mL culture flask for differential adhesion for 50 min. Then, the non-adherent cells were cultured for 12 h in an incubator at 37°C with 5% CO₂. The medium was changed after the cells had adhered and again every 3 days. On the 9th day, the culture was shaken at 250 rpm for 18 h to remove the microglia and oligodendrocytes into the upper medium layer. Finally, astrocytes with a purity of more than 95% were obtained after 3 days of culture. After CGS21680 (100 nM), SCH58261 (50 nM) (18), Colivelin (10 nM) (19), and Stattic (50 nM) (20) had been added separately, the primary astrocytes were cultured for 72 h under mild oxygen-glucose deprivation (mOGD) conditions; that is, in low-glucose Dulbecco’s modified Eagle’s medium/F12 medium (1 g/L) in a hypoxia incubator (3% O₂, 5% CO₂, 92% N₂). Subsequently the expression levels of the target factors were determined.

The animal tissue specimens and cell samples were fixed with 4% paraformaldehyde for 30 min, permeated with 0.1% Triton X-100 for 2 h, and blocked with 5% goat serum for 30 min (all carried out at ambient temperature). Then, the samples were incubated overnight at 4°C with primary antibodies against the following proteins: glial fibrillary acidic protein (GFAP) (1:1000 dilution; ab4674, abcam), STAT3 (1:100; Cat. No. 9145, Cell Signaling Technology (CST), Danvers, MA, USA), and YKL-40 (1:50; ab180569, abcam). Then, fluorescein-labeled secondary antibody and DAPI (4′,6-diamidino-2-phenylindole) stain solutions were added under dark conditions. Finally, the expression of the target proteins were detected under a confocal laser scanning microscope.

Total proteins were extracted from the cells with RIPA lysis buffer, and the protein concentration was determined using a bicinchoninic acid kit. Then, the proteins were denatured by boiling the sample for 5 min. Thereafter, the proteins were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis at a constant voltage of 80 V for 100 min, with the sample loading volume calculated on the basis of a 20 μg loading amount. After electrophoresis, the protein bands were electrotransferred to a polyvinylidene difluoride membrane at a constant current of 200 mA for 100 min. Subsequently, the membrane was blocked for 2 h at ambient temperature with 5% bovine serum albumin and then incubated overnight at 4°C with primary antibodies against GFAP (1:10,000 dilution; abcam), STAT3 (1:1000; Cat. No. 12640, CST), phospho-STAT3 (1:1000; Cat. No. 9145, CST), and YKL-40 (1:500; abcam). This was followed by incubation with the secondary antibody for 1 h at ambient temperature. Finally, a gel imaging system was used to visualize the proteins.

Total RNA was extracted using the TRIzol reagent. Then, after the concentration and purity of the sample had been verified, it was subjected to polymerase chain reaction (PCR) quantitation using a TaKaRa quantitative PCR kit (TaKaRa, Bio, Shiga, Japan) according to the manufacturer’s instructions. The reverse transcription conditions were 42°C for 15 min and 85°C for 5 s, and the amplification conditions were 40 cycles of 95°C for 30 s, 95°C for 5 s, and 60°C for 30 s. The results were visualized using a gel imaging system. The forward and reverse primers for YKL-40 and STAT3 were 5ʹ-CCGTTCCTGCGTTCTTAT-3ʹ and 5ʹ-ACTGGTTGCCCTTGGTAG-3ʹ, 5ʹ-GGAGGAGTTGCAGCAAAAAG-3ʹ and 5ʹ-TGTGTTTGTGCCCAGAATGT5ʹ, respectively, and were designed and synthesized by Sangon Biotech (Shanghai) Co., Ltd. (Shanghai, China).

After processing of the astrocytes, the enzyme-linked immunosorbent assay (ELISA) kit from eBioscience (San Diego, CA, USA) was used to detect the levels of inflammatory cytokines in the cell culture supernatant. Standards were prepared according to the kit instructions, and the supernatant samples were then added for incubation at 37°C for 30 min. After repeated washes with phosphate-buffered saline, the enzyme-labeled reagent was added for incubation at 37°C for 30 min. Finally, the stop solution was added to terminate the reaction, and the absorbance of the solution in each well was measured at 450 nm using a microplate reader.

SPSS 22.0 was used for all data analyses and all data are presented as the mean ± standard deviation (SD). One-way analysis of variance was applied for comparing the data of multiple groups, and Tukey’s test was used for comparisons between groups. Differences with a P value of less than 0.05 were considered statistically significant.

To verify the effects of ADORA2A, we administered an agonist and an inhibitor of the receptor to mice with surgically induced CCH to evaluate the degree of white matter injury. The results of immunofluorescence ( Figure 1 ) showed that the MBP absorbance was lower in the CCH group than that in the sham group and gradually decreased with prolongation of the ischemia time, confirming that CCH caused white matter injury in a time-dependent manner. Compared with that in the CCH group, the MBP absorbance was significantly lower in the SCH58261 group but significantly higher in the CGS21680 group, indicating that the activation of ADORA2A could inhibit the white matter injury induced by CCH. The MBP absorbance in the CGS21680 (KO) group was also lower than that in the CCH group, further confirming the protective role of ADORA2A against CCH-induced white matter injury.

To investigate how ADORA2A activation results in the reduction of white matter injury, we used immunofluorescence and PCR assays to detect the expression level of YKL-40 in the brain ( Figure 2A ). The immunofluorescence results showed that YKL-40 was barely expressed in the sham group, whereas its expression level was increased in the CCH group. Moreover, YKL-40 was mainly co-localized with GFAP, indicating that it was predominantly expressed in the astrocytes. Compared with the GFAP and YKL-40 levels in the CCH group, the expression in the SCH58261 group were significantly higher, whereas those in the CGS21680 group were significantly lower, indicating that the activation of ADORA2A could inhibit the activation of astrocytes and reduce the expression of YKL-40 ( Figures 2B–D ). The expression of GFAP and YKL-40 in the CGS21680 (KO) group were significantly higher than those in the CGS21680 (WT) group, further confirming that knockout of the ADORA2A gene promoted the CCH-induced activation of astrocytes and their expression of YKL-40. The PCR results showed the same expression trends for these target genes ( Figure 2E ). Taken together, these results verified that the activation of ADORA2A could inhibit astrocyte activation and decrease their expression of YKL-40.

To further evaluate the mechanism through which ADORA2A regulated YKL-40 expression, we evaluated the expression of STAT3 after CCH in mice. The immunofluorescence results showed the expression of STAT3 was higher in the CCH group than that in the sham group, suggesting that STAT3 was involved in the CCH-induced white matter lesions ( Figure 3A ). Compared with the GFAP and STAT3 levels in the CCH group, the expression in the SCH58261 group were significantly higher and those in the CGS21680 group were significantly lower, indicating that ADORA2A prevented the expression of STAT3 mediated by activation of the astrocytes ( Figures 3B–D ). The expression of STAT3 in the CGS21680 (KO) group was significantly higher than that in the CGS21680 (WT) group, further confirming that ADORA2A inhibited the expression of STAT3 in astrocytes.

Interestingly, the expression trend of STAT3 was consistent with that of YKL-40, suggesting that there was a certain causal relationship between them. Therefore, we used a STAT3 inhibitor (Stattic) and a STAT3 activator (Colivelin) on the WT mice to further evaluate these regulatory effects on YKL-40 expression. The western blot results revealed that the expression of YKL-40 was increased in the Colivelin group and decreased in the Stattic group relative to that in the CCH group, confirming that STAT3 could promote the expression of YKL-40 ( Figures 3E, F ). The expression of YKL-40 was significantly different between the CGS21680 and CGS21680+Colivelin groups, as were the levels between the SCH58261 and the SCH58261+Stattic groups, indicating that the activation of STAT3 could reverse the inhibitory effect of ADORA2A on YKL-40 expression. These results clearly demonstrate that ADORA2A reduces the synthesis of YKL-40 by inhibiting the expression of STAT3.

We used primary astrocytes in vitro to further verify the effect of ADORA2A on the STAT3/YKL-40 axis. The immunofluorescence results showed that after the mOGD treatment, astrocytes were activated and the expression of STAT3 and YKL-40 were increased, confirming that the two proteins were involved in the activation of astrocytes caused by chronic ischemia ( Figure 4A ). However, after activating ADORA2A with CGS21680 in the astrocytes, the expression of STAT3 and YKL-40 were decreased. By contrast, the inhibition of ADORA2A with SCH58261 in astrocytes led to increased expression of STAT3 and YKL-40. The same results were observed in the western blot ( Figures 4B–D ) and PCR ( Figures 4E, F ) experiments. These findings verified that ADORA2A regulated the STAT3/YKL-40 axis in astrocytes.

After inhibited ADORA2A with SCH58261, the astrocyte bodies became swollen and their axons were elongated and interwoven into a network, suggesting that the cells had been further activated ( Figure 5 ). Additionally, the expression of STAT3 and YKL-40 were increased. By contrast, astrocyte activation was significantly suppressed with CGS21680 intervention, further confirming that ADORA2A inhibited astrocyte activation. Moreover, compared with that in the SCH58261 group, the expression of YKL-40 was decreased by the further inhibition of STAT3 (SCH58261+ Stattic group). By contrast, the further activation of STAT3 (CGS21680 + Colivelin group) reversed the inhibitory effect of adenosine ADORA2A on YKL-40 synthesis that was observed in the CGS21680 group. These results verified that ADORA2A could downregulate YKL-40 synthesis by suppressing STAT3 expression, thereby inhibiting the activation of astrocytes.

Furthermore, we used ELISA to measure the astrocyte-mediated inflammatory response regulated by ADORA2A. The mOGD treatment promoted the IL-1β, tumor necrosis factor-alpha (TNF-α), and IL-6 secretion in the supernatant of the astrocyte, indicating that the low-oxygen conditions had activated the astrocytes and induced them to release inflammatory cytokines for mediating inflammation ( Figure 6 ). The expression of these cytokines in the SCH58261 and CGS21680 groups were significantly different from those in the CCH group and indicated that ADORA2A could inhibit astrocyte-mediated inflammation. The further inhibition of STAT3 (SCH58261 + Stattic group) reduced the cytokine secretion when compared with that in the SCH58261 group. By contrast, the further activation of STAT3 (CGS21680 + Colivelin group) reversed the inhibitory effect of ADORA2A on inflammation that was seen in the CGS21680 group. These results provided further evidence that ADORA2A inhibits astrocyte-mediated inflammation through STAT3/YKL-40 axis.