Forty-five healthy adult male Sprague-Dawley rats, aged 8 weeks and weighting 275 - 325 g, were provided by Jinan Pengyue Laboratory Animal Breeding Co. China, Ltd. [SCXK (Lu) 2014 0007]. The animals were housed individually in plastic boxes at 23˚C (±0.5˚C), drinking and eating freely. The Animal Ethics Committee of Qingdao University (Qingdao, China) approved all protocols for the Care and Use of Laboratory Animals.

45 SD rats were randomly divided into three groups, 15 in each. They were Control group, Model group (PTX) and Treatment group (PTX + HBOT), respectly. Rats in the treatment group and model group were intraperitoneally injected with paclitaxel 2 mg/kg/ml (paclitaxel dissolved in ethanol:cremophor EL:saline = 1:1:4) on the 1st, 3rd, 5th and 7th days of the experiment, and a stable peripheral neuralgia model could be established on the 14th day [13] . The control group was given intraperitoneal injection of solvent 1 ml/kg (ethanol:cremophor EL:saline = 1:1:4).

An animal HBOT chamber (YSD-I; Yantai Binglun Oxygen Industry Co., Ltd., Shandong, China) was utilized for HBOT treatment. 100% pure oxygen was filled with cabin, and quick lime was placed to further absorb CO₂ and water vapor. Animals in the Treatment group were then placed into the chamber. The pressure within was gradually increased at a rate of 0.1 ATA/ min until 2.5 ATA, and was maintained for one hour. At the end of the therapy, the pressure was gradually decompressed to atmospheric pressure at a rate of 0.1 ATA/min [10] . HBOT treatment lasted for 7 consecutive days (day 15 - day 21). Animals in model group were allowed to stay in the chamber without any treatment.

After treatment, an additional 7 days (day 22 - day 28) were to observe long-term effect on anti-allodynic and then rats were sacrificed for subsequent tests.

Rat mechanical withdrawal threshold (MWT, Mechanical Withdraw Threshold) was tested using Von Frey fiber (Stoelting Co.) by Up and Down method [14] . Stimulating the plantar surface of the hind paw for 3 - 5 s with an interval of 30 s, a withdrawal response was described as positive and turned to a lower force. In contrast, a higher force was conducted when a negative response occurred. 1 g, 1.4 g, 2 g, 4 g, 6 g, 8 g, 10 g, 15 g, eight Von Fery fibers were selected according to preliminary experiment, which started from 4 g and cut off at 15 g. The rats were tested 6 times after the first positive response, and finally the MWT (g) was calculated based on the arrangement of the positive and negative patterns. The MWT values were measured before modeling and during the treatment (Day 15, Day 17, Day 19, Day 21) and after treatment for observing long-term efficacy (Day 24, Day 28).

To examine astrocyte profiles on the long-term antiallodynic effect, five rats in each group were anesthetized with 10% chloral hydrate at 28^th day after behavioral tests, then the spinal cord L4-L6 were taken and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PBS), pH 7.4 overnight at 4˚C. After fixation, 5 - 7 μm sections of spinal cord were cut and washed in PBS and then incubated with 3% H₂O₂ to block peroxidase. They were washed in PBS three times and pre-incubated for 30 min at room temperature with 10% goat serum in PBS containing 0.7% Triton X-100. The sections were incubated overnight at 4˚C with rabbit anti-glial fibrillary acidic protein (GFAP) primary antibody [1:200] (Absin CO., China) and then after rinsing in PBS (3 × 10 min), incubated with secondary HRP-conjugated antibody for 30 min at room temperature. The substrate 3,30-diaminobenzidine (DAB) was applied for 5 min, and then were counterstained with hematoxylin. The signals of each group were normalized to Control group as fold increase for further statistical analysis using Image-Pro Plus Version 6.0.

To examine the expression levels of related proteins on the long-term antiallodynic effect, the spinal cords were taken from five rats as described above and kept fresh on the ice. Proteins were extracted by RIPA lysis buffer (p0013C, Beyotime, China), and then heated at 95˚C with loading buffer for 5 minutes for denaturation. 20 μg denaturated proteins for each were separated by 10% SDS-PAGE and then transferred onto a polyvinylidene difluoride (PVDF) membrane. After blocking for 1 hour at room temperature in Tris-buffered saline containing 5% non-fat milk, the membranes were incubated overnight at 4˚C with the primary antibody Rabbit anti-GFAP (Glial fibrillary acidic protein) polyclonal antibody (1:500; absin131715, Absin CO., China), Rabbit Anti-TNF alpha Polyclonal Antibody (1:500; absin123966, Absin CO., China), Rabbit Anti-IL-1 Beta Polyclonal Antibody (1:500; absin126104, Absin CO., China), Rabbit anti-β-actin monoclonal antibody (1:1000; absin132001, Absin CO., China). Following washes in TBST buffer, HRP-conjugated secondary antibody (1:2000; absin20002, Absin CO., China) was used to detect primary antibodies at room temperature for 2 hours. Peroxidase activity was visualized by ECL plus reagent and capture images. The intensity signals of each protein were normalized by that of β-actin and changes between groups were expressed as fold increase of other groups versus the control group.

To examine the expression levels of related mRNA on the long-term antiallodynic effect, the spinal cords were taken from five rats as described above and kept fresh on the ice. Total RNA was extracted according to the instructions of Trizol (R1100, Solarbio CO.) and the cDNA were obtained according to the instructions of the Reverse Transcription Kit (RP1105, Solarbio CO.). SYBR Green PCR kit (SR1120, Solarbio CO.) was used for PCR amplification and quantitative analysis. The common amplification conditions were set as: 95˚C 30 s, 60˚C 60 s, 72˚C 2 min, 45 cycles. The specificity of primers was observed by dissolution curve, and the CT values were recorded. Statistical analysis was carried out by using 2 − Δ Δ c t [15] , ΔCT = CT (target gene) - CT (GAPDH), ΔΔCT = ΔCT (target group) - ΔCt (mean value of ΔCT in control group for describing changes between groups as fold increase ). All ΔΔCT converted to 2 − Δ Δ c t for subsequent statistical analysis.

Primer [12] [16] :

Rat GAPDH

sense: 5’-ATGACTCTACCCACGGCAAG-3’

anti-sense: 5’CA TACTCTGCACCAGCATCTC-3’

Rat GFAP

sense: 5’-GAGTCCACAACCATCCTTCTGAG-3’

anti-sense: 5’-ACACCAGGCTGCTTGAACAC-3’

Rat IL-1β

sense: 5’-TGTGATGTTC CCATTAGAC-3’

anti-sense: 5’-AATACCACTTG TTGGCTTA-3’

Rat TNF-a

sense: 5’-CCACGCTCTTCTGTCTACTG-3’

anti-sense: 5’-GCTACGGGCTTGTCACTC-3’

Changes among groups in behavioral analysis and weight changes were compared using two-way repeated-measures analysis of variance (ANOVA) with post hoc Bonferroni tests. One-way ANOVA with post hoc Bonferroni correction was used to identify differences among groups in Immunohistochemistry and real-time RT-PCR. Data are presented as means ± standard errors of the mean. P values < 0.05 were considered significant. For real-time RT-PCR, P values < 0.01 were considered significant.

Body weight changes did not differ between paclitaxel- or solvent-treated animals (P > 0.05 Figure 1). There were no hair loss and diarrhea in all groups of rats during the experiment. It shows that the HBOT, paclitaxel and paclitaxel solvents have no influence on the health of rats under this experimental design.

On the 14th day after intraperitoneal injection of paclitaxel, the MWT value of the Control group was significantly lower than that of the Model group (P < 0.001, Day 14, Figure 2), indicating that the paclitaxel can cause significant hyperalgesia. The MWT value of the Treatment group had an upward trend and from the 19th day (5th day of treatment) statistical difference had been observed (P < 0.001, Day 19, Figure 2). The value maintained the level of the 19th day until the end of the treatment (P < 0.001, Day 21, Figure 2). After treatment, the value became to decrease, but still maintained a relative high level compared with Model group (P < 0.001, Day 24 Day 28, Figure 2), indicating that the analgesic effect of HBOT was sustainable after the cessation of treatment.

The expression of GFAP in the Model group was significantly higher than that in the Control group (P < 0.01, Figure 3; P < 0.01, Figure 4(a)), suggesting that paclitaxel can activate astrocytes. The Treatment group was significantly lower than that in the Model group (P < 0.01, Figure 3; P < 0.05, Figure 4(a)), indicating that HBOT inhibited the activation of astrocytes. The expression of mRNA levels of GFAP were no significant differences between groups (P = 0.0467, Figure 5(a), size of test p < 0.01).

The expression of IL-1β and TNF-α in the Model group was significantly higher than that in the Control group (P < 0.001, Figure 4(b) and Figure 4(c)), indicating that paclitaxel can induce neuroinflammation. The expression of IL-1β and TNF-α in the Treatment group was significantly lower than that in the Model group (P < 0.01, Figure 4(b) and Figure 4(c)), indicating that HBOT inhibited neuroinflammation. The expression of mRNA levels of IL-1β and TNF-α was significantly increased in the model group than that of Control group (P < 0.001, Figure 5(b) and Figure 5(c)), indicating that paclitaxel up-regulated the mRNA levels of IL-1β and TNF-α, resulting in a long-term neuroinflammation. The mRNA levels in Treatment group was significantly lower than that in the Model group (P < 0.01, Figure 5(b) and Figure 5(c)), indicating that HBOT could reduce the mRNA levels of IL-1β and TNF-α producing a long-term analgesic effect.