Sprague-Dawley rats (adult male), weighting 200–220 g were used (SLRC LABORATORY ANIMAL, Shanghai, China). All procedures were approved by the Experimental Animal Care and Use Committee of Shanghai University (No. 201506-766). Rats were housed at animal facility with controlled temperature (22–24°C) and photoperiod (12 h light/12 h dark) with freely access to water and food. Before experiments, animals were fasted for 24 h.

GABA_BR1(D-2) (sc-16608), NOX4 (H-300) (sc-30141), and CGRP (sc-8856) were all purchased from Santa Cruz Biotechnology (Santa Cruz Biotechnology, Santa Cruz, CA,USA). Phospho-Ser923 GABA_B R1 antibody was from ABGENT (WuXiAppTec, Shanghai, China). Anti-NOX2/gp91phox antibody [EPR6991] (ab129068) was from abcam (San Franciso, CA, USA). Goat anti-mouse IgG (H+L) Cy3 was from abcam (San Franciso, CA, USA). Anti-IL-1 beta antibody ab9722 was from abcam (San Franciso, CA, USA). Superoxide dismutase (SOD) assay lists was from Nanjing Jiancheng Biological Engineering Research Institute (Nanjing, China).

Chloral hydrate (450 mg/Kg) was used to anesthetize rats before experiment, and anesthetized rats then were fixed onto a stereotaxic apparatus. The coordinate for stimulating the FN (Paxinos et al., 2009) is AP 11.6 mm, LR 1.0 mm, and H 5.6 mm. L-Glu (3 μg in a volume of 0.5 μL saline) was microinjected 15 min before GI-R via a channula connected to a microsyringe with a polyethylene tube. The microinjection was performed for 3 min, and the injection cannula was left in situ for another 10 min to prevent backflow.

The location of the LHA was determined according to The Rat Brain in Sereotaxic Coordinates (Paxinos et al., 2009): LHA, anteroposterior (AP) 2.0 mm, left-right (LR) 1.5 mm, height (H) 8.3–8.4 mm. Baclofen (0.5, 1.5, 3.0 μg) in a volume of 0.5 μL saline and CGP35348 1.125 μg in a volume of 0.5 μL saline, were microinjected into the LHA via a cannula connected to a microsyringe with a polyethylene tube. The microinjection was performed for 3 min, and the injection cannula was left in situ for another 10 min to prevent backflow. Baclofen and CGP35348 were microinjected into the LHA 15 min before GI-R respectively.

GI-R was performed following L-Glu, Baclofen or CGP35348 microinjection as previously reported (Du et al., 2013). Briefly, the abdominal cavity was cut open for celiac artery and its adjacent tissues isolation. To induce ischemia, the celiac artery was clamped with a small vascular clamp for 30 min. The clamp was then removed to allow reperfusion for 30 min, 1, 2, 4, 8, 16 h. The rats were euthanized under deep anesthesia after each experiment. The rats' stomachs and brains were removed rapidly. The stomachs were opened along the greater curvature, and the gastric mucosa was grossly examined carefully for ulcers. Part of the brains were stored at −80°C for Western blotting, PCR and real-time PCR analysis. The other part brains were immersed in 4% paraformaldehyde for 48 h, and then immersed in 20% sucrose solution for 1 day, at last immersed in 30% for 1 day, sliced to 30 μm thick, mounted on glass slides for the immunofluorescent assay and ROS test.

The gastric mucosal injury was assessed as described previously (Zhang et al., 2002) with modification. Briefly, the stomach was spread out, and the gastric mucosal injury index (GMII) was counted. This index is based on a cumulative-length scale: an individual lesion limited to within the mucosal epithelium (including pinpoint erosions, ulcers, and hemorrhagic spots) is scored according to its length, for example, (1), value ≤1 mm; (2), value between 1 and 2 mm; (3), value between 2 and 3 mm. Additionally, For lesions >1 mm wide, the lesion score was doubled. Gastric mucosal injury index is thus the sum of the scores of all lesions, the index was determined by an investigator who was blind to the treatments to avoid researcher bias.

Immunofluorescence staining was performed as described previously (Zhang et al., 2011; Yang et al., 2012; Xu et al., 2013). The histological sections were localized in the LHA. Brain sections were permeabilized for 30 min with 0.1% Trion-X100 in PBS, then washed with PBS, then brain sections were incubated with primary antibody against GABA_BR1 (1:200, Santa Cruz Biotechnology) for 2 h at room temperature. After four times washes with PBS, brain sections were incubated with secondary antibodies coupled to Cy3 (1:500, Abcam Inc) for 1 h at room temperature. After washed with PBS, then incubated with 1% DAPI in PBS for 5 min. After four washes with PBS, brain sections were analyzed by a Nikon SA Microphot microscope.

Total protein from LHA and gastric mucosa tissues was extracted and protein concentration in cell lysates was determined using a BCA assay kit (Beyotime Shanghai China). Sixty micrograms of protein, obtained from each sample, were separated using SDS-PAGE (10% gel). Samples were then transferred to polyvinylidene fluoride membrane (Immobulon-P 0.45 μm, Millipore Germany). Five percentage of skim milk was used to block the membranes, which were then incubated overnight at 4°C with the primary antibodies GABA_BR1 (1:300); CGRP (1:200); NOX4 (1:200); Phospho-Ser923 GABA_BR1 Antibody (1:300); Anti-NOX2/gp91phox antibody (1:300); β-actin (1:1,000, rabbit, Santa Cruz biotechnology); Anti-IL-1beta antibody ab9722 (1:500). After washed with PBST, secondary antibodies (1:10,000, horseradish peroxidase-linked anti-rabbit or anti-goat, Santa Cruz biotechnology) was applied to the membranes for 1 h at room temperature. ECL (Immobulon, Millipore Germany) and Image Develop were used to visualize immunoblots.

TRIzol reagent (Takara technology) was used to extract the total RNA from the LHA nuclei issue. Purification and reverse transcription RNA was performed as follow: first-strand cDNA was synthesized from 3 μg total RNA using TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech). The sequences of GABA_BR primer used and the predicted amplification sizes are as follows: forward: CATCAACTTCCTGCCTGTG, reverse: GTGTCCATATCCGTCCAG, 247 bp.

Purification and reverse transcription RNA was performed as described above. Real time PCR was performed in a total volume of 50 μL with 1 μL reverse transcribed cDNA, 3 μL of each primer in hot start buffer, 0.5 μL hot start Taq DNA polymerase (Takara technology), 5 μL 10×LA TaqBuffer (Mg²⁺ Plus), and 8 μL dNTP Mixture. The sequences of GABA_BR and β-Actin primers used and the amplification sizes are as follows, GABA_BR forward: GGAAGGTGGCATCAGGTA, reverse: CATAGTCCAC AGGCAGGAA, 115 bp; β-Actin forward: GTACCCCATTGAAC ACGG, reverse: TGTGGTGCCAAATCTTCTC, 80 bp.

The LHA tissue was harvested from the rat brain and washed with ice-cold PBS. Subsequently, the LHA tissue was homogenized using western blot lysis buffer at 0–4°C, following which the LHA homogenate was centrifuged at 2,500 × g for 30 min. The supernatants were harvested and stored at −80°C until SOD activity assays was performed. The protein concentrations were determined by the BCA assay kit (Beyotime Shanghai China). A Xanthine/Xanthine Oxidase method was used to determine SOD activity, by measuring the absorbance value at a wavelength of 450 nm. A SOD Assay kit was used to measure SOD activity, according to the manufacturer's protocol (Nanjing Jiancheng Bioengineering Co., Ltd., Nanjing, China). Total SOD activities were expressed in uints/mg of protein.

The reactive oxygen species were detected on frozen section by ROS Fluorescent Probe-DHE (Vigorous Biotechnology, Beijing, China) recommended dilution of 1:100 (w/v) in 1% PBST for 1 h at room temperature and observed by fluorescence microscopy.

Data were analyzed using one-way ANOVA and LSD-test. Prism5 software was used (GraphPad Software, San Diego, CA, USA) to determine statistical difference between groups and statistical significance was accepted as P < 0.05. Data are presented as mean ± SD.

At first, we sought to determine whether BABA_B receptor agonist Baclofen (0.5, 1.5, 3.0 μg) exerts its protective effect on GI-R injury. We found that Baclofen dose-dependently attenuated GI-R injury (Figure 1, H–J). Furthermore, the gastric mucosal injury index (GMII) was time-dependently changes and the GMII peaked in the reperfusion 1 h (Figure 1, A–G).

As our previous reports, stimulation of FN with L-Glu attenuated the gastric ischemia-reperfusion injury in rats. Here, we want to investigate the effects of GABA_B receptor in LHA on GI-R injury. Microinjection of baclofen, a GABA_B receptor agonist or antagonist CGP35348 into the LHA. We found that unilateral microinjection of GABA_B receptor agonist baclofen into the LHA attenuated GI-R injury. However, microinjection of GABA_B receptor antagonist CGP38348 aggravated the GI-R injury (Figure 2). This result indicates that GABA_BR in the LHA plays a critical role in protecting the GI-R injury.

In order to detect whether the GABA_BR expressed in the LHA and to furtherly investigate the effects of baclofen and CGP35348 on GABA_BR expression, immunofluorescence technology was adopted to detect the GABA_B receptor. As shown in Figure 3, compared with the GI-R group, microinjection of L-Glu into the FN or GABA_BR agonist baclofen into the LHA up-regulated the GABA_BR expression. However, microinjection GABA_BR antagonist CGP35348 into the LHA down-regulated the GABA_BR expression.

Next, the expression of GABA_BR in the LHA was detected by western blot assay. As Figure 4 showed, microinjection of L-Glu into the FN or baclofen into the LHA significantly increased the expression of GABA_BR in the LHA, but microinjection CGP35348 decreased the expression of GABA_BR.

P-GABA_BR was detected by western blot assay. As shown in Figure 5, compared with the GI-R group, L-Glu into the FN or baclofen into the LHA lead the P-GABA_BR significantly decreased, P < 0.001. The level of P-GABA_BR is also quite low in the microinjection CGP35348 group, this may be related with the lower expression of GABA_BR. The expression levels of P-GABA_BR was opposed with GABA_BR.

To visualize the expression of GABA_BR in the LHA, we perform PCR and quantitative real-time PCR experiments. As shown in Figure 6, the amplification of GABA_B receptors from rats' brain cDNA revealed that all transcripts present in the LHA. Relative abundance of GABA_B receptors in the normal group was deviated from others and reached a peak in the Baclofen treatment group. But CGP35348 treatment inhibited the transcript. In addition, microinjection of L-Glu into the FN or baclofen into the LHA increase the GABA_BR-mRNA expression, compared with the GI-R group. The levels of GABA_BR-mRNA were identified with the levels of GABA_BR protein in each corresponding groups.

L-Glu microinjected into the FN or baclofen microinjected into the LHA attenuated the GI-R injury, but CGP35348 microinjection into the LHA aggravate the GI-R injury. Meanwhile, as shown in Figures 4, 5, L-Glu microinjected into the FN or baclofen microinjected into the LHA enhanced the expression levels of GABA_BR protein or the GABA_BR-mRNA. According to these results, we want to investigated how the baclofen or the CGP35348 influences gastric ischemia-reperfusion injury, and whether baclofen or CGP35348 affect the NOX2/NOX4 expression and ROS production?

As shown in Figure 7, compared with the GI-R group, the expression of NOX2 and NOX4 were all up-regulated in different drugs treatment groups. The gastric ischemia-reperfusion induces up-regulation of NOX2 and NOX4, baclofen or CGP35348 treatment doesn't reverse this changes. This result indicates that the protection of microinjection of GABA_BR agonist in the LHA in GI-R injury isn't mediated via oxidative stress pathway.

Total SOD activity was considered as an indication of the anti-oxidative properties of cells. The ROS fluorescent Probe-DHE was detected on the frozen section. As shown in Figure 8A, compared to the GI-R group, there was no change in different drugs treatment groups. Compared to the GI-R group, SOD activity in the LHA had little significantly change in the other groups, as shown in Figure 8B.

The above studies indicated that the protective effect of Baclofen was not mediated by the oxidative stress pathway in brain, then we furtherly investigated the relations between GABA_B receptor and gastric mucosa injury. We detected IL-1β, NOX2, and NOX4 in the gastric mucosa. As shown in Figure 9, compared to B group, the expression of IL-1β decreased in L-Glu treatment group, ^***P < 0.001 and D group, ^###P < 0.001; the expression level of IL-1β increased in E group, ^▴▴P < 0.01.

Compared to B group, the expression level of NOX2 decreased in C group, ^**P < 0.001 and D group, ^##P < 0.001; the expression level of NOX2 increased in E group, ^▴▴P < 0.01, as shown in Figure 10. Compared to B group, the expression of NOX4 decreased in C group, ^***P < 0.001 and D group, ^###P < 0.001; the expression level of NOX4 increased in E group, ^▴▴P < 0.01, as shown in Figure 10.

Base on our previous report that CGRP could attenuate the GI-R injury, we determined whether Baclofen in the LHA can induce the release of CGRP in the gastric tissue. As shown in Figure 11, compared with the GI-R group, the expression level of CGRP increased LHA+GI-R+Baclofen group, ^*P < 0.01.

Finally, to confirm the CGRP attenuated the GI-R injury via anti-oxidant pathway in gastric mucosa, CGRP was administrated (iv) and levels of superoxide anion radical generating enzymes NOX2/4 were measured by immunoblot. As shown in Figure 12, compared with the GI-R group, CGRP attenuated the expression of NOX2 and NOX4, ^*P < 0.001.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.