Hyperbaric Oxygen Suppressed Tumor Progression through Improvement of Tumor Hypoxia and Induction of Tumor Apoptosis in Lung Cancer

Tumor cells have long term been recognized as a relative contraindication to hyperbaric oxygen treatment (HBOT) since HBOT might enhance progressive cancer growth. However, in an oxygen decit condition, tumor cells are more progressive and have the potentials to be metastatic. HBOT increasing in oxygen partial pressure may benet tumor suppression. In this study, we investigated the effects of HBOT on solid tumors, such as lung cancer. Non-small cell human lung carcinoma A549-cell-transferred severe combined immunodeciency mice (SCID) mice were selected as an in vivo model to detect the potential mechanism of HBOT in lung tumors. HBOT not only improved tumor hypoxia but also suppressed tumor growth in murine xenograft tumor models. In vitro, HBOT suppressed the growth of A549 cells in a time-dependent manner and immediately downregulated the expression of p53 protein after HBOT in A549 cells. Our results demonstrated that HBOT improved tissue vasculogenesis, tumor hypoxia and potentially target apoptosis to lung cancer cells in murine xenograft tumor models. HBOT will merit further cancer therapy as an adjuvant treatment for lung cancer.


Introduction
Tumor hypoxia has been a major problem for cancer therapy [1,2]. Hypoxia can increase tumor resistance to chemotherapy, radiation and lead to malignant progression and even metastasis [3,4].
Hypoxia or some signaling alterations of the hypoxic cascades could enhance the intact neovascularization, which facilitates tumor metastasis [5,6].
Hyperbaric Oxygen Treatment (HBOT) is a medical treatment using 100% oxygen administered at a greater than normal atmospheric pressure to patients. Since cellular and vascular proliferation is promoted by HBOT in an ischemic wound [7,8], practitioners of hyperbaric medicine have concerned about the effect of HBOT on cancer growth [9]. However, more and more evidence suggests a neutral effect of HBOT on malignancy [10][11][12]. HBOT working as an adjuvant to promote cellular and vascular proliferation might have the same effect in a tumor. However, the nature physiology between tumor growth and wound healing is very different by the way of cancer growth, transformation and metastases. Such differences will result in different impacts on HBOT. A major difference pointed out by Growther et al [13] suggested macrophages are the major healing factors in angiogenesis, while tumor macrophages contributed as part as an angiogenesis factor in a tumor environment [14,15]. The optimal oxygen tensions of 30 to 40 mmHg increased by HBOT can stimulate collagen synthesis and hydroxylation in wounds, but is not working to promote cell proliferation in cancer.
Our previous experiments demonstrated that exposure to HBO attenuated the severity of disease progression in autoimmune NZB/W F1 mice [16]; proposing targeted apoptosis to hyper-proliferating lymphocytes. Moreover, cells of hematopoietic origin have a lower threshold to oxidative stress induced by HBOT and HBO-induced apoptosis of hematopoietic derived cancer cells was through the intracellular accumulation of H 2 O 2 and O 2 .as well as the involvement of phosphorylation of p38 MAPK [17].
Lung cancer is responsible for more than one million deaths annually and is diagnosed with the most popular cancer in the world. The main types of lung cancer are small-cell lung cancer (SCLC), and non- HBOT suppressed tumor growth.
On the observation that HBOT could decrease the proliferation rate of A549 cells, we established A549 tumor transferred-SCID mice xenograft model for the potential use of HBOT (2.5 ATA, 90 min) in cancer therapy. A549-tumor transferred SCID mice were either non-exposed or exposed to HBO for 2.5 ATA 90 min 5 times in a week beginning at the 45 th date after tumor injection and were monitored for the tumor volume twice in a week. The tumor volume of human lung carcinoma A549 cells was increased as a linear curve after 20 days of transfer for both control and HBO-treated mice( Figure 1A).Only the mice with tumor volume larger than 500mm 3 were selected to HBOT. Our data showed that HBOT effectively  HBOT induced apoptosis in A549 tumor transferred SCID mice.
Histopathological ndings of tumor lesion were quite similar at the 0 th date after HBO exposure between control and HBOT groups. The area of the necrotic eld is not observed at the 0 th date after HBOT; however necrotic eld was more prominent at the14 th and 28 th date after HBOT. Ghost-like shadow of cells was recognized in the necrotic eld as "coagulation necrosis". Tumor sections of HBOT-14 th and 28 th date showed zonal in ammatory cells and degenerated cells with nuclear debris were aggregated in the border of coagulation necrosis ( Figure 3A). Immunostaining of cleaved-caspase-3 was demonstrated in HBO-treated mice after 14 days of HBOT ( Figure 3B) compared with control mice ( Figure 3C). CD31 was signi cantly increased after HBOT.
Platelet endothelial cell adhesion molecule (PECAM-1/CD31), is a ligand for CD38 and plays a role in angiogenesis. CD31 is highly expressed in endothelial cells and the expressions of vascular endothelial growth factor (VEGF) and CD31 have recently been implicated in tumor angiogenesis. HBOT did not change the expression of VEGF but signi cantly increased the expression of CD31 after 14 days and 28 days of HBOT (*P< 0.05)( Figure 4, Table 1). HBOT decreased p53 protein and increased HIF-1a protein in A549 cells, not Beas-2B cells.
We examined the effects of HBOT on A549 cells compared with an immortalized human bronchial epithelial cell line, Beas-2B, followed by the cell-based HBOT protocol ( Figure 5A). HBOT suppressed the growth of A549 cells, not Beas-2B cells, in a time-dependent manner as measured in Figure 5B. From western blotting analysis, p53 proteins in A549 cells were rst downregulated by HBOT and then rebounded to basal level after 6 hours of HBOT, but not Beas-2B cells ( Figure 5C). There's no signi cant difference in the expression of p53 mRNA after HBOT between A549 and Beas-2B cells ( Figure 5D).The LC3BII/I ratio was increased by the HBOT and then declined to basal level after HBOT in Beas-2B cells ( Figure 5C).
HBOT-reducedp53 protein could be rescued by a proteasome degradation inhibitor, but not an autophagy inhibitor in A549 cells.
The downregulation of p53 protein of A549 cells upon HBOT was consistent at a continuous 3-day exposure; p53 protein was immediately downregulated at the rst moment of HBOT exposure and back to the normal after 20 h exposure ( Figure 6). Hypoxia-inducible factor 1 alpha (HIF-1a) protein expression was signi cantly decreased by HBOT after exposure at the rst interval and rebounded to basal level both in A549 and Beas-2B cells ( Figure 6); however, the apoptotic biomarkers of cleaved PARP and caspase 3 remained unchanged in both A549 and Beas-2B cells ( Figure 6). There was no obvious increaseofsubG1 phase by ow cytometry analysis ( Figure 6). With the cytosolic and nuclear fractions, we found that the effect of HBOT on p53proteinsoccurred in the cytosol and nucleus of A549 cells ( Figure 7). We tested whether a proteasome degradation inhibitor, MG132 or the autophagy inhibitor 3-methyladenine (3-MA), could rescue the effect of HBOT on p53 and HIF-1a proteins. Our data showed that MG132 could rescue p53 and HIF-1a proteins in A549-HBOT cells ( Figure 7). The effect of MG132 on HIF-1a protein was dependent on cell-context and HBOT. In A549 cells, MG132 not only stabilized basal HIF-1a protein, but also further enhanced HIF-1a protein in A549-HBOT cells. However, the HIF-1a protein in MG132-treated Beas-2 cells was still decreased under HBOT condition. 3-MA could suppress the induction of autophagy by HBOT in Beas-2B cells, suggesting that the autophagy did not affect p53 protein stability in A549 cells by HBOT (Figure 7).

Discussion
In this study, HBO was used to treat mice transferred with A549 human lung carcinoma in the manner of 2.5 ATA/90 min daily for 2 weeks after 45 days of tumor establishment. HBOT in a cycle of 14 and 28 days both showed signi cant improvement for tumor hypoxia environment and suppressed tumor growth compared with control mice. Interestingly, tumor vascularity detected by the expressions of CD31 was signi cantly increased after 14 and 28 days of HBOT; however, the expression of VEGF did not change as measured in semi-quantitative IHC staining analysis. This nding indicated that the improvement of cancer hypoxia and cancer vascularity by HBOT did not promote cancer growth. For tumor angiogenesis, the new blood vessels are formed when the basement membrane of existing blood vessels is broken and while the vascular basement membrane has been breached, endothelial cells will be divided to form vessels and start to grow. VEGF appears to be the rst to induce endothelial cell mitosis in tumor angiogenesis [21,22]. In contrast, CD31, which is a mitogenic factor in wound healing, reacts more sensitive to oxygen tension [23]. HBOT signi cantly increased the expression of CD31 after 14 days of HBOT. Meanwhile, the volume of A549 tumor cells was not growing as increasing expressions of CD31.
To evaluate the direct effect of HBOT in A549 cells, the decrease of p53 and HIF-1a proteins was involved in the ubiquitin-dependent proteasome degradation pathway via the proteasome degradation inhibitor MG132 in A549 cells. The extra effect of MG132 on p53 and HIF-1a proteins were observed by HBOT in A549 cells. In a normal lung epithelial cell line, we also observed the decrease of HIF-1a protein with the HBOT, even in the presence of MG132. The involvement of autophagy was veri ed via the increasing LC3BII/I ratio and its speci c inhibitor, 3-MA, in Beas-2 cells, but not in A549 cells. Our ndings suggest that HBOT induced tumor suppression in A549-tumor transferred SCID mice might be through modi cation of tumor microenvironment rather than induction of autophagy in tumor cells themselves.
Malignant tumors have been recognized as a relative contraindication to HBOT . It was concerned that HBOT might have cancer growth-enhancing effects. Our ndings indicated that the improvement of tumor hypoxia and tumor vascularity by HBOT did not promote tumor growth but inhibit tumor development, which provided evidence for the potential use of HBOT on tumor malignancy. Although the mechanism of HBOT on tumor suppression is not clear, our experiments argue that HBOT results in tumor apoptosis. Moreover, HBOT can increase the formation of reactive oxygen species (ROS) and induce apoptosis of tumor cells [24,25].

Injection of A549 tumors
Mice were anesthetized by intra-peritoneum (i.p.) injection of pentobarbital: 40mg/kg body weight. The abdomen was sheared and sterilized with 2-isopropanol (70%).A549 cells were diluted to 6 x 10 6 cells in a volume of 0.2ml RPMI 1640 and transferred to each recipient at the neck site by subcutaneous injection. The diameter of tumor mass was monitored two times per week and only mice with tumor volume larger than 500mm 3 were selected for further experiments.
Hyperbaric oxygen treatment (HBOT) A549-tumor transferred SCID mice were either non-exposed or exposed to HBO in a hyperbaric animal chamber (98% O 2 , 2% CO 2 at 2.5 ATA; 0.5 ATA/min to a pressure of 2.5 ATA) (Longshin Gas Ltd. Taipei, Taiwan, ROC)for 90 min once a day over two weeks beginning at the 45 th date after tumor injection. Cells in groups were exposed to HBOT (98% O2, 2% CO2 at 3.5 ATA ) (Longshin Gas Ltd., Taipei, Taiwan, ROC)in a hyperbaric chamber for 90 min. Control cultures for each experiment were placed in an incubator at 37 •C, 21% O2,5% CO2 at 1 ATA. This hyperbaric chamber was maintained at a concentration of 95% oxygen, and CO 2 was exhausted at a rate of 10-12 L/min. Reaching the end of the treatment, the chamber was slowly decompressed at 0.5 ATA/min pressure.

Tumor volume
We measured tumor volume two times in a week and calculated the tumor size as length (L) x width (W) x height (H).

Detection of tumor hypoxia
Hypoxyprobe TM -1 kit (Pimonidazole Hydrochloride) (Chemicon International) was used to measure tissue hypoxia in solid tumors. A dose of 60mg/kg bodyweight of Hypoxyprobe TM -1 kit was injected intraperitoneally to A549-transferred SCID mice and mice were sacri ced after 1h. Tumor mass was removed and embedded as OCT/liquid nitrogen cryostat section.

Tumor histopathology
Tumor cells were removed from mice for either frozen or para n sections. For frozen sections, tissue cells were snap-frozen in dry ice-cold 2-methylbutane and embedded in TISSUE-TEC (Miles, Elkhart, IN, USA). Freshly cut sections (5 mm) were mounted on clean glass slides coated with poly-L-lysine (Sigma Chemical Co, USA). They were further rapidly air-dried and stored at -80 o C until used for immunohistochemical staining. The xed sections were incubated with normal goat serum diluted 1:5 in PBS for 15 min at room temperature to block nonspeci c staining. Sections were later incubated with goat anti-mouse IgG primary antibody (anti-CD31; BD Pharmingen, San Diego, CA, USA)in PBS for 1 h at room temperature and washed with PBS for 5 min with gentle shaking. After washing in PBS, they were interacted with secondary antibodies and developed with DAKO Kit-DAB chromogen for 6 min. The slides were washed with Q water and incubated with hemataxylin for the staining of the nucleus. The sections were immersed with 75%-95%-100% alcohol and xylene before covered with a coverslip. The processing, embedding and sectioning of para n blocks were performed in Dr. Koichi Tsuneyama's pathology laboratory and the para n sections were depara nization and re-hydration before immunohistochemical staining.
x (Abcam, UK). Immunoreactive proteins were visualized using horseradish peroxidase-linked secondary antibodies and further with ECL (Enhanced-chemiluminescence

Statistical analysis
Unpaired t-test, one-way ANOVA and Nonparametric Kruskal-Wallis test were used to compare differences between means of groups and statistical analysis was performed using SPSS. One-Dscan analysis software was used to analyze the expression contents of proteins in western blotting.

Declarations
Authors' contributions  Figure 1 A549-tumor transferred SCID mice were either non-exposed or exposed to HBO2(98% O2, 2% CO2 at 2.5 ATAfor2.5 ATA 90 min 5 times in a week beginning at the 45th date after tumor injection and were monitored for the tumor volume twice in a week. Tumor volume was calculated as length (L) x width (W)  group (B) and control group (C). Note that, in HBOT group, many cleaved-caspase-3 positive apoptotic cells were located in the border between coagulation necrosis and viable tumor cells.

Figure 4
Comparative analysis of immunohistochemical CD31 antigen expressions from groups of mice.
Signi cantly increase of CD31 was observed after 14 days and 28 days of HBOT exposure lysates were subject to the (C) Western analysis for antibodies against p53, LC3B, and control protein ACTN and (D) RT-PCR analysis for p53 mRNA and control GAPDH mRNA. The grouping of gels/blots was cropped from different parts of the different gels and the full-length gels and blots were included in the supplementary gure 5. Please note that there was no difference in the expression patterns for both shorter and longer exposure (Supplementary gure 5).

Figure 6
The HBOT effect on stress proteins on A549 and Beas-2 cell lines. The HBOT A549 and Beas-2 cells were treated with 3.5 ATA for 90 mins and then lysates were subject to the Western analysis for antibodies against p53, LC3B, HIF-1α, PARP, Caspase 3, and control protein ACTN. The grouping of gels/blots was cropped from different parts of the different gels and the full-length gels and blots were included in the supplementary gure 6. Please note that there was no difference in the expression patterns for both shorter and longer exposure (Supplementary gure 6).
γH2A.x, Nrf2, PARP, and control protein ACTN. The grouping of gels/blots was cropped from different parts of the different gels and the full-length gels and blots were included in the supplementary gure 7.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.