Rat EGC/PK060399egfr (CRL-2690™) and human intestinal epithelial cells Caco-2 (HTB-37™) were obtained from the American Type Culture Collection. EGC/PK060399egfr and Caco-2 cells were grown in high glucose DMEM and MEM, respectively, supplemented with 10% FCS, 2 mM l-glutamine, and 100 U/100 μg/ml penicillin–streptomycin. Cells were incubated in a 5% CO₂ humidified incubator at 37°C. The EGC-Caco-2 co-culture system was established as described previously by our laboratory (Xiao et al., 2014). Caco-2 cells were seeded on Millicell®filters (0.4 μm pore diameter; Millipore; Billerica, MA) at a density of 5 × 10⁴ cells/cm² for up to 4–5 days. EGCs were seeded at an equal density in 6 or 24 well tissue culture plates to avoid any possible direct cell contact with Caco-2 cells. During the co-culture period, half of the culture medium in the apical and basal compartments was changed daily.

Global A_2AR homozygous KO mice (A_2AR^−/− mice) with C57BL/6J background were provided by Dr. Yuanguo Zhou (Research Institute of Surgery, Daping Hospital, Army Medical University, Chongqing, China). Specific pathogen-free wild-type (WT) C57BL/6J mice were purchased from the Laboratory Animal Center of the Army Medical University. All mice were housed and maintained in laminar flow cabinets under specific pathogen-free conditions.

For hypoxia experiments, cells were subjected to hypoxia in a CO₂ incubator (Forma® Series II Water Jacketed CO₂ Incubators; Thermo Scientific) with 94% nitrogen, 5% CO₂, and 1% oxygen and incubated at 37°C for 6 h. Re-oxygenation was initiated by replacing the media and exposing the cell monolayers to 37°C plus 5% CO₂for 1 h. Control cells were maintained at 37°C in an atmosphere with 5% CO₂.

Male mice (8–10 weeks old) were fasted for 12 h and were free to drink water prior to surgery. The animals were intraperitoneally injected with 40 mg/kg of pentobarbital anesthesia and an aseptic laparotomies dioventral line was placed. The following specific surgical procedures were performed as previously described (Xiao et al., 2014).

Cells and tissues were lyzed in cold RIPA buffer for 30 min and centrifuged at 13,000× g for 30 min at 4°C. Protein concentration was determined using a BCA assay reagent (Beyotime). The primary antibodies used were rabbit anti-ZO-1 (1:800), rabbit anti-Occludin (1:1000), mouse anti-A_2AR (1:500), rabbit anti-PKCα (1:1000), rabbit anti-Na+/K+ ATPase (1:1000), and rabbit anti-GAPDH (1:1,000). Protein expression was measured in optical density units and normalized to GAPDH expression.

The small intestine tissues were embedded with OCT compound (Tissue-Tek, Sakura Finetek, Torrance, CA, United States). Consecutive frozen sections (5 μm in thickness) were obtained and fixed in 4% paraformaldehyde for 20 min at room temperature. After 30 min pre-incubation with a blocking solution containing 5% bovine serum albumin, sections were incubated overnight at 4°C with primary antibody against GFAP (Abcam), A_2AR (Abcam), or ZO-1 (Abcam). After washing in PBS, sections were incubated with fluorescence-conjugated secondary antibodies at 37°C for 1 h. After washing in PBS, sections were incubated with DAPI nuclear stain solution for 5 min. All images were obtained using a TCS-SP5 confocal microscope (Leica, Germany).

Cells were harvested and lyzed in standard immunoprecipitation (IP) buffer containing either 1% 3-[(3-cholamidopropyl) dimethylammonio] propanesulfonic acid (Chaps) or 1% Triton X-100 (for ERAD substrates), or 2% digitonin (for ERAD machinery) for 1 h on ice. Cells were centrifuged at 16,000 × g for 10 min, and the supernatant was used for immunoprecipitation experiments. Co-IP was performed using protein A-agarose beads (EMD Millipore) with anti-A_2AR (Santa Cruz Biotechnology, Inc.), anti-mGluR5 antibodies (Cell signaling), or anti-D2R antibodies (Santa Cruz Biotechnology, Inc.) following the usual method.

The TER of cells was determined via a Millipore electric resistance system (ERS-2; Millipore). Caco-2 cell monolayers were grown in Millicell inserts (0.33 cm² area, 0.4 μm pore diameter, and 6.5 mm diameter) and the culture medium was replaced before TER measurement. To calculate the actual resistance of the cell monolayer, the mean resistance of filters without cells was subtracted from the monolayer measurement, and the difference between the filter and monolayer areas was corrected.

The intestinal mucosa TER was measured by Ussing chambers (Physiologic Instruments, San Diego, CA). The excised intestinal tissues were bathed in 5 ml Krebs buffer (110.0 mM NaCl, 3.0 mMCaCl₂, 5.5 mM KCl, 1.4 mM KH₂PO₄, 29.0 mM NaHCO₃, and 1.2 mM MgCl₂, pH 7.4)on both the mucosal and serosal sides. The TER was measured as previously described (Liu et al., 2018).

All experimental data are shown as the Mean ± SD. Statistical significance was determined by unpaired two-tailed Student t test analysis using GraphPad Prism version 7.0 software (San Diego, CA). If not otherwise stated, all experiments included three independent replications in triplicate. p < 0.05 was considered statistically significant.

To study the effect of A_2AR on EGC, we first examined the expression of A_2AR in different pathological conditions. Lipopolysaccharide (LPS) and hypoxia treatments were used to stimulate EGCs in vitro. As shown in Figure 1A, A_2AR expression dramatically increased in EGCs following LPS and hypoxia stimulation, with the effect of hypoxia being obvious than that of LPS. Therefore, hypoxia treatment was used to study the effect of A_2AR. As demonstrated previously, GFAP is a specific marker of activated glial cells (Xiao et al., 2014). A_2AR and GFAP immunofluorescent staining colocalization were used to observe changes in A_2AR expression in EGCs. Acute IR-treated mice showed a moderate decrease in GFAP-positive intestinal EGCs compared to sham-treated mice (Figure 1B). However, the A_2AR levels increased significantly in GFAP-positive EGCs after IR treatment. Together, these results indicate that hypoxia stimulation can activate the A_2AR-mediated signaling pathway in mucosal EGCs in the intestine.

To further explore the role of A_2AR in IEB modulation under acute hypoxia stimulation, we used an A_2AR agonist and inhibitor separately in an in vitro EGC-Caco-2 co-culture system. The tight junctions (TJs) are primary determinants of IEB function (Lee, 2015). As shown in Figure 2A, western blot analysis revealed an almost 50% drop in ZO-1 and occludin expression in the hypoxia group and a similar reduction on ZO-1 and occludin expression in the A_2AR inhibitor ZM241385 group. However, the A_2AR agonist CGS21680 significantly prevented hypoxia-induced TJs destruction. Further study with immunocytochemistry (ICC) confirmed the western blot results (Figure 2B). Additionally, TER measurement analysis produced similar results. CGS21680 pretreatment effectively blocked TER decrease under hypoxia stimulation when compared to the ZM241385 pretreatment group (528.6 ± 18.11 Ω cm² vs. 289.6 ± 12.63 Ω cm² for the CGS21680 and ZM241385 pretreatment groups, respectively) (Figure 2C). Together, these results showed that A_2AR plays a protective role in the EGC barrier-protecting effect on the IEC monolayer under hypoxia stimulation.

To further confirm the role of A_2AR in the barrier protective of EGCs during hypoxia stimulation, A_2AR KO mice were treated with IR. Western blot analysis of ZO-1 and occludin expression was assessed in the small intestine (Figure 3A). Acute IR treatment led to a substantial decrease in ZO-1 and occludin expression in A_2AR KO mice compared to WT mice. Immunofluorescence analysis also revealed a similar reduction in ZO-1 and occludin expression in the intestinal mucosa after IR stimulation (Figure 3B). The functional impact of A_2AR knockdown on tight junctions in the small intestine was further evaluated by determining the TER value using Ussing chambers. As shown in Figure 3E, intestinal I/R caused a marked TER decrease in A_2AR KO mice (78.63 ± 3.407 Ω cm²) compared with WT mice (94.5 ± 3.151 Ω cm²). Histological examination of intestinal tissues revealed that IR-treated A_2AR KO mice showed more increased intestinal villus fracturing and epithelial removal than did WT mice (Figures 3C,D). Together, these results strongly suggest that A_2AR has a significant protective effect in IR-induced IEB injury.

There are synergistic interactions between A_2AR, mGluR5, and the dopamine D2 receptor (D2R) in central nervous system (CNS) related diseases (Ferraro et al., 2012; Fernández-Dueñas et al., 2013). However, whether this relationship exists in the enteric nervous system has yet to be determined. It has been suggested that in micro glial cells, A_2AR combines with D2R in low glutamate concentration and combines with mGluR5 in high glutamate concentration (Dai et al., 2010; Beggiato et al., 2016). Therefore, we used high glutamate concentrations as a positive control. We studied the relationship among them under hypoxia in EGCs. When exploring the effect of oxygen concentration, the band corresponding to D2R was coimmunoprecipitated by anti-A_2AR antibodies, and an A_2AR band was coimmunoprecipitated by anti-D2R antibodies. Together, these results indicate that A_2AR and D2R interact in normoxic but not in hypoxic conditions (Figure 4A). However, an opposite relationship was observed between A_2AR and mGluR5. As shown in Figure 4B, the band corresponding to mGluR5 was coimmunoprecipitated by anti-A_2AR antibodies, and an A_2AR band was also coimmunoprecipitated by anti-mGluR5 antibodies. These results indicate that A_2AR and D2R interact in hypoxia but not in normoxia, demonstrating that A_2AR interacts with D2R under normoxic conditions and interacts with mGluR5 under hypoxic conditions.

To confirm that A_2AR regulates the IEB via a mGluR5-dependent pathway, we tested the responsiveness of the IEB to 100 μM of the MPEP selective mGluR5 antagonist under hypoxia. Western blot analysis of ZO-1 and occludin expression revealed that treatment with MPEP significantly inhibited CGS21680-mediated activation of ZO-1 and occludin expression after hypoxia induced damage (Figure 4C). Similar results were observed in TER measurements: CGS21680 pretreatment effectively prevented the decrease of TER from hypoxia stimulation, while, MPEP blocked CGS21680-induced potentiation (454.3 ± 15.7 Ω cm² vs. 391 ± 8.441 Ω cm²) (Figure 4D). Together, these results indicate that the protective effect of A_2ARon the IEB after hypoxia stimulation is dependent on mGluR5.

The PKCα signaling pathway is associated with an A_2AR–mGluR5 interaction-associated proinflammatory effect (Dai et al., 2013; Li et al., 2017). Therefore, we next tested whether PKCα is required for A_2AR-mediated IEB protection. PKC family isoforms can translocate to multiple subcellular localizations in response to hypoxia in different cell lines (Yu et al., 2015). Consistent with this, western blots showed increased PKCα expression in the membrane after hypoxia stimulation in EGCs (Figure 5A). These results were confirmed by immunocytochemistry analysis showing that PKCα translocates from the cytoplasm to the cell membrane in EGCs after hypoxia stimulation (Figure 5B). We then used PKCα inhibitor, chelerythrine chloride, to observe whether the protective effect of CGS21680 on IEB was affected. As shown in Figure 5C, chelerythrine chloride significantly reduced ZO-1 and occludin protein levels in CGS21680-pretreated EGCs. As expected, chelerythrine chloride pretreatment also blocked CGS21680-promoted TER (533 ± 24.2 Ω cm² vs. 390.9 ± 24.14 Ω cm²) (Figure 5D). Together, these results suggest that A_2AR exerts its protective effects on IEB via the PKCα signaling pathway.