Hydrogen Gas Attenuates Hypoxic-Ischemic Brain Injury via Regulation of the MAPK/HO-1/PGC-1a Pathway in Neonatal Rats

A total of 320 neonatal rats were randomly divided into the following eight groups (n = 40/group): HIBI, HIBI+H₂-30 min, HIBI+H₂-60 min, HIBI+H₂-90 min, HIBI+H₂-90 min (12 h), HIBI+H₂-90 min (24 h), sham surgery, and control. Animals in the HIBI and all HIBI+H₂ treatment groups were rats with induced HIBI. After establishing the HIBI model, neonatal rats in the HIBI+H₂-30 min, HIBI+H₂-60 min, and HIBI+H₂-90 min groups were immediately placed in a box for 30, 60, or 90 min of H₂ inhalation, respectively. This was followed by twice daily H₂ inhalation (30, 60, or 90 min each time) over the next 3 days (7 : 00–9 : 00 am and 7 : 00–9 : 00 pm). Animals in the HIBI+H₂-90 min (12 h) and HIBI+H₂-90 min (24 h) groups underwent 3% H₂ inhalation for 90 min at 12 and 24 h, respectively, after induction of HI. This was followed by twice daily H₂ inhalation for 90 min for 3 days. Animals in the sham surgery group underwent left carotid artery isolation without ligation and hypoxia induction. Animals in the control group were healthy neonatal rats without induction of HIBI. Detection of early neurological reflexes was performed ~24 h after HIBI induction, and then some neonatal rats were subjected to triphenyltetrazolium chloride (TTC) staining (n = 8/group). At 96 h after inducing HIBI, brain samples from neonatal rats were collected for western blotting, paraffin-embedded, and sectioned for terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and immunofluorescence staining. In addition, some of the neonatal rats (n = 8/group) in adulthood (postnatal days 70–74) underwent testing in a Morris water maze (Figure 1). The study was performed in a blinded and randomized matter with respect to treatment administration and histological and functional assessments.

Neonatal Sprague-Dawley rats (males and females, 7 days old, 12–18 g), were provided by the Department of Laboratory Animals, Central South University (Hunan Province, China). All pups were housed with their mothers before and after the experiments in a barrier environment (all measurement indicators were in conformity with the national standard “Experimental Animal Environment and Facilities GB14925, 2010”). All of the experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (National Institute of Health Publication No. 80-23, revised 1996), and approved by the Center for Medical Ethics of the Third Xiangya Hospital (Hunan, China). Sincere efforts were made to minimize animal suffering and to reduce the number of animals used.

Neonatal rats at 7 days of age were anesthetized with isoflurane inhalation. This was followed by making a 0.7 cm incision to the left of the cervical midline to separate the left common carotid artery, which was severed after bilateral ligation. Surgical suture 6-0 was used to close the incision. The entire surgical procedure was completed within 8 min. Neonatal rats recovered in the same cage as their mother for 2 h after surgery. Next, they were placed in an anoxic tank (Gaoke Gas Technology Co. Ltd., Changsha, China) containing 8% oxygen and 92% nitrogen at 37°C to induce hypoxia for 100 min. After hypoxic exposure, neonatal rats were allowed to recover or underwent H₂ inhalation. The inhaled gas for H₂ intervention was a mixture of H₂ and air. H₂ was generated by electrolysis of water in the MHG-2000 Hydrogen Generator (MiZ Co. Ltd., Kanagawa, Japan) at a concentration of 3%.

The infarct volume was determined by TTC staining. Using this method, the brain sections were prepared as follows. First, the brains were removed and frozen at -20°C for 10 min. Next, consecutive 2 mm coronal sections were obtained by slicing the brains with a Brain Matrix (ASI Instruments, Warren, MI, USA). The subsequent incubation of the sections was performed in a dark environment with a 30 min immersion in 2% TTC solution at 37°C. TTC stained normal areas of brain deep red but did not stain the infarcted tissue. Infarction volumes were measured and analyzed with ImageJ software (version 1.61; NIH Image, Bethesda, MD, USA).

After the last H₂ treatment, rats were perfused with cold 0.1 M phosphate-buffered saline (PBS, pH 7.4) through the left ventricle. The brains were fixed in 4% paraformaldehyde overnight. Fixed brain tissues were further processed by dehydration in graded ethanol prior to paraffin embedding. Finally, the blocks were serially cut into 5 μm thick sections and stored at 4°C. After deparaffinization, the sections were pretreated within antigen unmasking solution (Vector Laboratories, Burlingame, CA, USA) for 15 min in the microwave oven (medium fire for 8 min, ceased fire for 8 min, and low fire for 7 min). After washing with PBS, the sections were quenched with spontaneous fluorescence quenching agent (Servicebio, Wuhan, China) for 5 min at room temperature. Nonspecific binding sites were blocked by incubating the sections in 5% bovine serum albumin for 30 min. This was followed by an overnight incubation at 4°C with the following primary antibodies: anti-glial fibrillary acidic protein (GFAP) (1 : 1000; Sigma, St. Louis, MO, USA), anti-ionized calcium-binding adaptor molecule 1 (1 : 1000; Wako, Osaka, Japan), anti-HMGB1 (1 : 500; Abcam, Cambridge, MA, USA), anti-TLR4 (1 : 100; Servicebio), anti-Nrf2 (1 : 100; Abcam), anti-HO-1 (1 : 100; Abcam), anti-BAX (1 : 100; Abcam), anti-Bcl2 (1 : 100; Abcam), anti-caspase3 (1 : 100; Servicebio), and anti-p-p38 MAPK (1 : 100, Abcam). Then the sections were immersed in PBS and incubated with the secondary antibody (Servicebio) for 50 min at room temperature. Finally, the sections were incubated for 10 min with DAPI. After washing with PBS, the sections were sealed with antifluorescence quenching sealant and scanned using Pannoramic MIDI (3DHISTECH Ltd., Budapest, Hungary). CaseViewer software was used to merge the stained images and integrate the fluorescence emission intensity.

Twenty-four hours (24 h) after establishing the HIBI model, early neurological reflexes [24] including righting, cliff aversion, and geotaxis reflexes were evaluated in neonatal rats as follows. (1) The righting reflex was assessed by placing each neonatal rat in a supine position and recording the time (in seconds) to roll over to a normal prone position. For each neonatal rat, the test was performed three times to obtain a mean value. (2) The cliff aversion reflex was assessed by dangling the upper limbs of each neonatal rat from a board edge and recording the time to turn 90° away from the cliff edge. The maximum duration of observation was 20 s. Neonatal rats that did not turn 90° within the observation period were recorded at a time of 20 s. (3) The geotaxis reflex was measured by placing the head of each neonatal rat downward on a 40° inclined plate and recording the time required for turning around (with the head facing upward and >90° rotation). The maximum duration of observation was 20 s. Neonatal rats that did not turn within the observation period were recorded at a time of 20 s.

The Morris water maze was used to study the effects on spatial learning and memory in rats [25, 26]. The test was performed at week 10 after establishing the HIBI model (postnatal days 70–74, representing rats in adulthood [27]). Rats were tested in a pool of 160 cm in diameter. The pool was filled with water at a depth of 40 cm. The rescue landing platform (8 cm in diameter) was submerged by 1 cm under the water surface in one of four quadrants. Water temperature was maintained at 23°C–24°C. Testing consisted of four consecutive training days during which four attempts (starts from four quadrants) were given to each rat to find the platform within 60 s. Each rat was given 20 min to rest between swimming trials. The training session occurred at the same time of day throughout the training and testing periods. A rat was placed in the water in the quadrant opposite to the landing, and the latency time (in seconds) spent by each animal to find the platform was recorded. If the rat was not able to locate and climb on the landing platform within the allotted time, the experimenter gently placed the animal on the platform and let the animal become accustomed to it for 20 s. On day 5 of probe training testing, the rescue platform was removed and the time spent in the quadrant where the landing used to be was recorded (allotted time 60 s). Only one attempt was given to each rat during the probe trial on day 5 of testing.

Rats were sacrificed by decapitation after the last H₂ treatment, and the brain was microdissected and stored at -80°C. NGF-differentiated PC12 cells were washed with PBS and lysed in RIPA buffer (Proteintech, Rosemont, PA, USA) for 30 min on ice. The protein concentration in the extracts was determined using the BCA assay (Wellbio, Changsha, China). Equal amounts of protein (30 μg) were separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride membranes (Millipore, Billerica, MA). Membranes were blocked in 5% skim milk powder dissolved in phosphate-bufferred saline with 0.5% Tween 20 in PBS (PBST) at 4°C overnight and then incubated with the following antibodies: anti-GFAP (1 : 1000; Abcam), anti-HMGB1 (1 : 1000; Abcam), anti-TLR4 (1 : 500; Proteintech, Rosemont, IL, USA), anti-phosphorylated nuclear factor kappa B (p-NF-κB) p65 (1 : 1000; Abcam), anti-caspase3 (1 : 1000; Abcam), anti-BAX (1 : 1000; Abcam), anti-Bcl-2 (1 : 500; Proteintech), anti-p-p38 MAPK (1 : 1000; Abcam), anti-p-ERK1/2 (1 : 1000; Abcam), anti-p-JNK antibody (1 : 1000; Abcam), anti-Nrf2 (1 : 1000, Abcam), anti-HO-1 (1 : 1000; Abcam), anti-PGC-1α (1 : 1000; Abcam), and anti-SIRT1 (1 : 1000; Abcam). β-Actin (Proteintech) was used as the gel loading control. Following overnight incubation, blots were washed three times with PBST and then incubated with the secondary antibody (IRDye 800CW Goat Anti-Rabbit IgG, 1 : 10000; Invitrogen, Carlsbad, CA, USA). Proteins were detected using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA). Band intensity was quantified using ImageJ software.

The PC12 cell line was obtained from American Type Culture Collection (Manassas, VA, USA) and maintained at 37°C in a humidified atmosphere containing 5% CO₂ in high-glucose Dulbecco's Modified Eagle's Medium supplemented with 10% heat-inactivated fetal calf serum, 10% heat-inactivated horse serum, 100 kU/L penicillin, and 100 mg/L streptomycin. For differentiation, PC12 cells were plated at a low density on collagen-coated plastic in DMEM plus 1% horse serum and NGF (50 ng/mL) for 7 days.

The cell culture was incubated at 37°C in a low oxygen incubator (1% O₂, 94% N₂, and 5% CO₂) for 6 h (OGD). Then, glucose-free DMEM was replaced with normal medium, followed by incubation for 3/6/24 h (OGD/R) in standard culture gas (N₂ as base gas, O₂ 21%, and CO₂ 5%). For the OGD/R+H2 group, cells were incubated for 3/6/24 h in standard culture gas with H₂ gas (N₂ as base gas, O₂ 21%, CO₂ 5%, and H₂ 2–3%). Immediately after inducing OGD/R or OGD/R+H2, cells were collected for western blotting, quantitative PCR (qPCR), fluorescence-activated cell sorting (FACS), and immunofluorescence staining.

All RNA was extracted from cells using a DNA/RNA Mini Kit (QIAGEN, Germantown, MD USA). Reverse transcription was done using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA USA). Quantitative measurements of target gene expression relative to 18 s RNA were performed in triplicate using the TaqMan Real-Time PCR assay following the manufacturer's recommendations in a Real-Time PCR system.

Cells were detached and collected into a 15 mL centrifuge tube, washed twice with ice-cold PBS, and centrifuged at 1000 rpm for 5 min to remove the supernatant. A volume of 0.1 mL binding buffer containing 0.14 M NaCl, 2.5 mM CaCl₂, and 0.01 M HEPES/NaOH pH 7.4 was added to resuspend the cells, followed by adding 5 μL Annexin V-FITC and 5 μL of 50 μg/mL propidium iodide (PI) staining reagents. After mixing homogeneously and reacting at 25°C for 15 min in the dark, the apoptotic cell population was analyzed by a flow cytometer.

Mean ± standard deviation was calculated for all parameters determined in this study. Repeated measures of two-way analysis of variance (ANOVA) were used for each parameter of the water maze task, followed by multiple t-test analyses. One-way ANOVA followed by Tukey's multiple-comparison and western blotting was used for histopathological parameters (volume and cell density). P < 0.05 was considered statistically significant.

As shown in Figures 2(a) and 2(c), TTC staining showed no obvious white infarcts in either the left or right cerebral hemisphere in the normal and sham surgery groups but showed obvious white infarcts in the left cerebral hemisphere of the HIBI group, with ligation of the left common carotid artery. The area of cerebral infarction was significantly reduced with an increase in duration of H₂ inhalation (^∗p < 0.05; ^∗∗p < 0.01). As shown in Figure 2(b), antibodies for two different biomarkers, neuronal nuclei (NeuN) and GFAP, were used in immunofluorescence staining to observe changes in neurons and astrocytes in the hippocampus. Compared with the normal and sham surgery groups, the immunoreactivity of NeuN was significantly reduced in the cortex and hippocampal CA3 regions of the HIBI group (^∗p < 0.05; ^∗∗p < 0.01; Figure 2(d)), which suggested that HIBI caused neuronal loss. In addition, the immunoreactivity of GFAP was significantly increased in the HIBI group (^∗p < 0.05; ^∗∗p < 0.01; Figure 2(e)), which suggested that astrocytes were activated. However, compared with the HIBI group, the immunoreactivity of NeuN was significantly increased in the cortex and hippocampal CA3 regions of the H₂ inhalation groups, especially in the H₂-90 min group, and the immunoreactivity of GFAP was significantly decreased in the H₂ inhalation groups. Thus, H₂ inhalation might protect neurons from injury and reduce activation of astrocytes. We also found that the decrease of neuronal loss was significantly associated with the duration and initiation time of H₂ inhalation. H₂ inhalation for 90 min immediately after HIBI had the most significant protective effect (^∗p < 0.05; ^∗∗p < 0.01), while 90 min H₂ inhalation 24 h after HIBI had no significant protective effect on neuronal injury but did reduce activation of astrocytes, which suggests that astrocytes may be involved in the processes of HIBI.

To evaluate the neuroprotective effects of H₂ on HI brain injury, we evaluated the righting, geotaxis, and cliff aversion reflexes of rats 24 h after HIBI. Neonatal rats in the HIBI group had a significantly slower righting reflex, and H₂ inhalation improved the righting reflex in this group. Times for the righting reflex in the neonatal HIBI rats were significantly shortened with earlier initiation and extended duration of H₂ inhalation (^∗p < 0.05, ^∗∗p < 0.01 vs. the sham surgery group ^∗∗∗p < 0.01; Figure 3(a)). In assessment of the geotaxis reflex, neonatal rats in the HIBI group took significantly longer to turn from facing downward on the inclined plate, with the time to turn being three times greater than that of the sham surgery group. H₂ intervention improved the geotaxis reflex (^∗p < 0.05, ^∗∗p < 0.01 vs. the sham surgery group ^∗∗∗p < 0.01; Figure 3(b)). However, no significant difference was found between the H₂-60 min and H₂-90 min groups (p > 0.05). Improvement in time for turning was less pronounced with later H₂ intervention in neonatal HIBI rats (Figure 3(b)). Assessment of the cliff aversion reflex showed results similar to those of the righting and geotaxis reflexes. Neonatal rats in the HIBI group had longer cliff avoidance times, and H₂ inhalation decreased cliff avoidance time in this group. However, significant improvement in cliff avoidance time was only found in the H₂-60 min and H₂-90 min groups (^∗∗p < 0.01 vs. the sham surgery group ^∗∗∗p < 0.01; Figure 3(c)).

The spatial memory performance was evaluated using the Morris water maze. Rats were subjected to repeated Morris water maze tests 10 weeks after HI exposure or sham operation. The results showed that neonatal HIBI induced significant long-lasting cognitive deficits throughout brain maturation and adulthood. As shown in Figures 4(a) and 4(b), the mean latency to find the platform declined progressively during the period of training days in all animals except the HIBI and H₂-90 min (24 h) groups. The latent period of the HIBI group was significantly longer than that of the control and sham surgery groups. H₂ inhalation significantly shortened the latent period. This effect was correlated with the duration of H₂ inhalation, and the H₂-90 min group had the most significant improvement among the H₂ intervention groups. Delays in initiation of H₂ intervention reduced the neuroprotective effects of H₂ against brain injury. The navigational memory of the rats was assessed by removing the escape platform on the fifth day of the Morris water maze test. Neonatal HIBI rats with H₂ intervention in different groups demonstrated better navigational memory than the HIBI group, and the H₂-60 min, H₂-90 min, and H₂-90 min (12 h) groups showed significant differences (^∗p < 0.05; ^∗∗p < 0.01; Figures 4(c) and 4(d)).

To explore the protective mechanism of H₂ on HIBI, we assessed the expression of HMGB1, TLR-4, and NF-κB-p65 in the hippocampus of neonatal rats. As shown in Figure 5, we found significantly enhanced expression of TLR-4 and GFAP in the HIBI group, suggesting that astrocytes were associated with enhanced TLR-4 signaling. In the HIBI+H₂ groups, TLR-4 and GFAP expression was lower compared with the HIBI group and the longer the H₂ inhalation, the lower the TLR-4 and GFAP expression. We also found that the HMGB1 expression was mainly localized in the nuclei. In the HIBI group, HMGB1 expression was significantly enhanced and detected outside the nuclei and GFAP expression was also enhanced. In the HIBI+H₂ groups, GFAP expression was reduced compared with the HIBI group and HMGB1 expression was confined to the nuclei, suggesting that the secretion of HMGB1 was reduced after H₂ inhalation (^∗∗p < 0.01; Figure 5).

As shown in Figure 6(a), TUNEL staining showed no detectable neuronal apoptosis in the sham surgery group but showed elevated apoptosis neurons in the cortex and hippocampal CA3 region of the HIBI group. The number of apoptotic neurons was significantly reduced in the H₂-90 min group. The Annexin-V/PI double staining assay further confirmed that H₂ attenuated apoptotic events in NGF-differentiated PC12 cells following OGD/R. Representative assay results are shown in Figure 6(b), and statistical analyses are summarized in Figure 6(c). Both early (Annexin V⁺/PI⁻) and late (Annexin V⁺/PI⁺) stages of apoptotic PC12 cells increased in the OGD/R group, while H₂ reduced the number of apoptotic PC12 cells (^∗p < 0.05; ^∗∗p < 0.01). Bcl-2 and BAX genes have been recognized as the two most important regulatory genes that are functionally antagonistic to each other in the regulatory processes of apoptosis. Bcl-2 inhibits apoptosis, while BAX promotes apoptosis. Caspase-3 is the most critical apoptosis executioner protease. In this study, the HIBI group showed downregulated Bcl-2 protein expression and upregulated BAX and caspase-3 protein expression (Figure 6(d)). Bcl-2 protein expression in the H₂ intervention groups was higher than that in the HIBI and OGD/R groups, while BAX and caspase-3 protein expression in the H₂ intervention groups was lower than that in the HIBI and OGD/R groups (^∗∗p < 0.01; Figures 6(e)–6(g)).

We examined the expression of p-p38 MAPK, HO-1, and Nrf2 in the cortex of HIBI rats. As shown in Figure 7(a), H₂ significantly enhanced the p-p38 MAPK, HO-1, and Nrf2 expression in the HIBD rat cortex. As shown in Figures 7(b) and 7(c), H₂ significantly enhanced the p-p38 MAPK, p-ERK1/2, p-JNK, HO-1, and Nrf2 expression in NGF-differentiated PC12 cells following OGD/R (^∗∗p < 0.01).

The effects of p38 MAPK, ERK, and JNK inhibition on the expression of HO-1 are shown in Figure 8. The p38 MAPK pathway inhibitor SB203580, ERK inhibitor U0126, and JNK inhibitor SP600125 significantly reduced OGD/R-induced injury, and H₂ plus OGD/R injury induced HO-1 expression (^∗∗p < 0.01).

The effects of HO-1 silencing on apoptosis and expression of PGC-1α and SIRT1 are shown in Figure 9. NGF-differentiated PC12 cells treated with HO-1 small interfering RNA (siRNA) showed increased apoptosis events following OGD/R injury, and H₂ did not attenuate apoptosis (Figures 9(a) and 9(b)). As shown in Figures 9(c)–9(i), HO-1 siRNA markedly decreased the expression of HO-1, PGC-1α, and SIRT1 induced by H₂ in NGF-differentiated PC12 cells following OGD/R (^∗p < 0.05; ^∗∗p < 0.01).

The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request.