Arachidonic Acid Metabolites of CYP450 Enzymes and HIF-1α Modulate Endothelium-Dependent Vasorelaxation in Sprague-Dawley Rats under Acute and Intermittent Hyperbaric Oxygenation

Acetylcholine-induced vasorelaxation (AChIR) and responses to reduced pO2 (hypoxia-induced relaxation (HIR), 0% O2) were assessed in vitro in aortic rings of healthy male Sprague-Dawley rats (N = 252) under hyperbaric (HBO2) protocols. The studied groups consisted of the CTRL group (untreated); the A-HBO2 group (single HBO2; 120 min of 100% O2 at 2.0 bars); the 24H-HBO2 group (examined 24 h after single exposure) and the 4D-HBO2 group (four consecutive days of single HBO2). AChIR, sensitivity to ACh and iNOS expression were decreased in the A-HBO2 group. HIR was prostanoid- and epoxyeicosatrienoic acid (EET)-mediated. HIF-1α expression was increased in the 24H-HBO2 and 4D-HBO2 groups. LW6 (HIF-1α inhibitor) decreased HIR in the 24H-HBO2 group. HBO2 affected the expression of COX-1 and COX-2. CYP2c11 expression was elevated in the 24H-HBO2 and 4D-HBO2 groups. Concentrations of arachidonic acid (AA) metabolites 14(15)-DiHET, 11(12)-DiHET and 8(9)-DiHET were increased in A-HBO2 and 24H-HBO2. An increased concentration of 8(9)-EET was observed in the A-HBO2 and 24h-HBO2 groups vs. the CTRL and 4D-HBO2 groups, and an increased concentration of 5(6)-DiHET was observed in the 24H-HBO2 group vs. the 4D-HBO2 group. The 20-HETE concentration was increased in the A-HBO2 group. All were determined by LC-MS/MS of the aorta. The results show that AChIR in all groups is mostly NO-dependent. HIR is undoubtedly mediated by the CYP450 enzymes’ metabolites of AA, whereas HIF-1α contributes to restored HIR. Vasoconstrictor metabolites of CYP450 enzymes contribute to attenuated AChIR and HIR in A-HBO2.


Introduction
Although the beneficial effects of hyperbaric oxygenation (HBO 2 ) on tissue perfusion are well documented, there is a paucity of knowledge about the mechanisms by which HBO 2 improves tissue oxygenation. It is known that various arachidonic acid metabolites (prostaglandins, epoxyeicosatrienoic acids (EETs), hydroxy-eicosatrienoic acids (HETEs)) and NO are of the utmost importance in mediating vascular reactivity to vasodilators and vasoconstrictors [1][2][3][4][5], including hypoxia and hyperoxia stimuli [6]. In conditions of reduced blood flow, the use of HBO 2 can significantly increase tissue oxygenation. Intermittent exposure to HBO 2 has been shown to cause activation of the CYP450 epoxygenase pathway in the macrovasculature of diabetic animals and to restore vasorelaxation of aortic rings in response to acetylcholine [7]. Furthermore, Kibel et al. (2012; showed an improved relaxation response to ANGII and ANG-(1-7) in healthy animals treated by intermittent HBO 2 due to activation of mechanisms related to CYP450 enzyme activation and EET synthesis [8,9]. However, the way in which acute or intermittent HBO 2 affects the mechanisms of vasodilation to physiological stimuli such as reduced pO 2 or acetylcholine (ACh) in healthy blood vessels is still not known. Since HBO 2 has been increasingly used both for health (e.g., sport) treatments and cosmetic purposes, understanding the mechanisms of action and consequences of the utilization of HBO 2 in healthy individuals is of primary importance.
Recently, we showed that increased superoxide production and an overall increase in oxidative stress impairs vasorelaxation in rats acutely exposed to HBO 2 , whereas intermittent exposure to HBO 2 increases the expression and activity of antioxidative enzymes, suggesting that intermittent HBO 2 induces antioxidative defense mechanisms underlying restored vasorelaxation [18]. Altogether, there is no clear answer as to which mediators are involved in the vasoconstriction or vasorelaxation observed in various HBO 2 protocols, because up to the present study, no measurements of vascular eicosanoids production have yet been performed.
The role of transcription factor hypoxia inducible factor 1 alpha (HIF-1α) in the effects of HBO 2 was described in studies on wound healing (diabetic foot, postiradiation injury). HIF-1α is involved in neovascularization, cell proliferation and increased mobilization of progenitor endothelial cells from bone marrow [19,20]. HBO 2 activates HIF-1α at several levels by increasing both HIF-1α stability and activity [21]. Hypoxia and reactive oxygen species (ROS)/ reactive nitrogen species (RNS) may stimulate HIF-1α stabilization, leading to the activation of hypoxia-induced cellular signaling pathways [22]. There is a limited knowledge on the role of HIF-1α on vascular reactivity in conditions of elevated oxidative stress and exposure to HBO 2 .
Taken together, the hypothesis of the present study was that HBO 2 changes the mechanisms of vascular relaxation and affects HIF-1α expression. Furthermore, HIF-1α is involved in vascular relaxation mechanisms, depending on the duration and frequency of HBO 2 exposure. Therefore, the purpose of this study was to investigate whether the mechanisms of ACh-and hypoxia-induced vascular relaxation in healthy rats and the expression of HIF-1α in aortic tissue were influenced by acute and intermittent HBO 2 . A further goal was to determine the involvement of metabolites of AA, produced in CYP450 enzymatic pathways, in vasorelaxation in different HBO 2 protocols.

Body Mass and Blood Pressure of Studied Groups
Body mass (g) of rats was not significantly different among examined groups. There was a significant decrease in mean arterial pressure in the A-HBO 2 group, compared to other groups (Table 1).  Figure 1a presents the results of isolated aortic ring baseline vasorelaxation in response to ACh (AChIR) in all experimental groups of rats. AChIR was significantly reduced in the A-HBO 2 group (the group receiving a single HBO 2 exposure) compared to all other groups of animals. A-HBO 2 rats also exhibited lower sensitivity to ACh compared to other groups of rats (presented by logEC50 in tables). HBO 2 significantly affected AChIR by reducing the inhibition of vasorelaxation caused by L-NAME compared to the control (CTRL) group, although sensitivity to ACh was not significantly different ( Figure 1b).

Acetylcholine-Induced Vasorelaxation (AChIR) of Isolated Rat Aortic Rings
The role of COX-1 and COX-2 metabolites was assessed with indomethacin. In the 4D-HBO 2 group (receiving four consecutive days of single HBO 2 exposure), AChIR was significantly improved after indomethacin compared to the other groups. In the 24H-HBO 2 group (examined 24 h after single exposure) indomethacin partially inhibited AChIR at a dose of 10 −7 mol L −1 compared to the CTRL group ( Figure 1c). Furthermore, MS-PPOH, an inhibitor of epoxygenase reactions significantly inhibited AChIR in the A-HBO 2 group compared to the CTRL and 24H-HBO 2 groups at ACh concentrations of 10 −9 -10 −7 mol L −1 , and compared to the 4D-HBO 2 group at 10 −7 mol L −1 . MS-PPOH partially inhibited AChIR at 10 −8 mol L −1 in the 4D-HBO 2 group versus the CTRL group ( Figure 1d). Figure 2 presents the mechanisms of AChIR in each experimental group of rats. Administration of L-NAME eliminated AChIR in the CTRL, A-HBO 2 and 24H-HBO 2 groups. In the 4D-HBO 2 group, the AChIR of the aortic rings was partially inhibited by the use of L-NAME and partially inhibited by MS-PPOH at ACh doses of 10 −6 mol L −1 and 10 −7 mol L −1 (Figure 2d).
The sensitivity to ACh (presented by logEC50 in the tables) decreased with L-NAME in the CTRL group. In the A-HBO 2 group (Figure 2b), indomethacin increased the sensitivity of aortic rings to ACh, whereas L-NAME and MS-PPOH decreased the sensitivity to ACh compared to the baseline response. In the 24H-HBO 2 group (Figure 2c), ACh sensitivity was mostly reduced by L-NAME compared to the baseline response. MS-PPOH and indomethacin also reduced the sensitivity to ACh compared to the CTRL group. In the 4D-HBO 2 group, ACh sensitivity was also decreased in the presence of L-NAME, and increased in the presence of indomethacin and MS-PPOH (Figure 2d).

Figure 1.
Ach-induced relaxation (AChIR) of isolated rat aorta rings in the CTRL, A-HBO2, 24H-HBO2 and 4D-HBO2 groups. Baseline relaxation to acetylcholine (a) and relaxation to acetylcholine in the presence of eNOS inhibitor L-NAME (b), COX inhibitor INDO (c) and a selective CYP450 epoxidase inhibitor, MS-PPOH (d) in CTRL, A-HBO2, 24H-HBO2 and 4D-HBO2 groups of rats. Results are presented as mean ± SD; N -number of aortic rings; p < 0.05. Figure 2 presents the mechanisms of AChIR in each experimental group of rats. Administration of L-NAME eliminated AChIR in the CTRL, A-HBO2 and 24H-HBO2 groups. In the 4D-HBO2 group, the AChIR of the aortic rings was partially inhibited by the use of L-NAME and partially inhibited by MS-PPOH at ACh doses of 10 −6 mol L −1 and 10 −7 mol L −1 (Figure 2d).
The sensitivity to ACh (presented by logEC50 in the tables) decreased with L-NAME in the CTRL group. In the A-HBO2 group (Figure 2b), indomethacin increased the sensitivity of aortic rings to ACh, whereas L-NAME and MS-PPOH decreased the sensitivity to ACh compared to the baseline response. In the 24H-HBO2 group (Figure 2c), ACh sensitivity was mostly reduced by L-NAME compared to the baseline response. MS-PPOH and indomethacin also reduced the sensitivity to ACh compared to the CTRL group. In the 4D-HBO2 group, ACh sensitivity was also decreased in the presence of L-NAME, and increased in the presence of indomethacin and MS-PPOH (Figure 2d). Results are presented as mean ± SD; N -number of aortic rings; p < 0.05.

Hypoxia-Induced Vasorelaxation of Isolated Rat Aortic Rings
The A-HBO2 group exhibited significantly decreased baseline vasorelaxation in response to hypoxia (hypoxia-induced relaxation (HIR)) compared to all other groups of rats. The 24H-HBO2 and 4D-HBO2 groups exhibited significantly enhanced HIR compared to the CTRL and A-HBO2 groups of rats (Figure 3a

Hypoxia-Induced Vasorelaxation of Isolated Rat Aortic Rings
The A-HBO 2 group exhibited significantly decreased baseline vasorelaxation in response to hypoxia (hypoxia-induced relaxation (HIR)) compared to all other groups of rats. The 24H-HBO 2 and 4D-HBO 2 groups exhibited significantly enhanced HIR compared to the CTRL and A-HBO 2 groups of rats ( Figure 3a compared to the baseline response. The use of L-NAME had no effect on the relaxation response to hypoxia in any of tested groups (Figure 3b- inhibitor) compared to the baseline response. The use of L-NAME had no effect on the relaxation response to hypoxia in any of tested groups (Figure 3b-e).

Relative mRNA and Protein Expression of Enzymes in Rat Aorta
Relative mRNA expression of COX-1 was significantly reduced only in the 24H-HBO2 group, compared to other groups (Table 2, Figure 4). mRNA expression of COX-2 was significantly reduced in the 4D-HBO2 group, whereas protein expression of COX-2 was significantly increased compared to controls (Table 2, Figure 5b). COX-2 mRNA was also significantly reduced in the 24H-HBO2 and 4D-HBO2 groups, compared to the A-HBO2 group. The relative mRNA expression of iNOS was significantly reduced in the A-HBO2 and 24H-HBO2 groups, compared to the control group, whereas eNOS expression was not significantly different among the groups (Table 2, Figure 4).

Relative mRNA and Protein Expression of Enzymes in Rat Aorta
Relative mRNA expression of COX-1 was significantly reduced only in the 24H-HBO 2 group, compared to other groups (Table 2, Figure 4). mRNA expression of COX-2 was significantly reduced in the 4D-HBO 2 group, whereas protein expression of COX-2 was significantly increased compared to controls (Table 2, Figure 5b). COX-2 mRNA was also significantly reduced in the 24H-HBO 2 and 4D-HBO 2 groups, compared to the A-HBO 2 group. The relative mRNA expression of iNOS was significantly reduced in the A-HBO 2 and 24H-HBO 2 groups, compared to the control group, whereas eNOS expression was not significantly different among the groups (Table 2, Figure 4). Table 2. Relative mRNA expression of cyclooxygenase 1 and 2 (COX 1 and COX 2), nitric oxide synthase (iNOS and eNOS), hypoxia inducible factor 1 alpha (HIF-1α), vascular endothelial growth factor (VEGF) and CYP2C11 in rat aorta normalized to HPRT housekeeping gene.

CTRL
A-HBO 2 24H-HBO 2 4D-HBO 2 Data are presented as mean ± SD (standard deviation); n = 6 (number of rats per group). * p < 0.05 compared to CTRL group; † p < 0.05 compared to A-HBO 2 group; ‡ p < 0.05 compared to 24H-HBO 2 group. Table 2. Relative mRNA expression of cyclooxygenase 1 and 2 (COX 1 and COX 2), nitric oxide synthase (iNOS and eNOS), hypoxia inducible factor 1 alpha (HIF-1α), vascular endothelial growth factor (VEGF) and CYP2C11 in rat aorta normalized to HPRT housekeeping gene. Data are presented as mean ± SD (standard deviation); n = 6 (number of rats per group). * p < 0.05 compared to CTRL group; † p < 0.05 compared to A-HBO2 group; ‡p < 0.05 compared to 24H-HBO2 group. Relative expression of CYP2c11 mRNA was significantly elevated in the 4D-HBO2 group, compared to other groups (Table 2), whereas CYP2c11 protein expression was significantly increased in 24H-HBO2 and 4D-HBO2 compared to the CTRL group ( Figure 5d). Other investigated proteins were similarly expressed among groups ( Figure 5). Relative expression of CYP2c11 mRNA was significantly elevated in the 4D-HBO 2 group, compared to other groups (Table 2), whereas CYP2c11 protein expression was significantly increased in 24H-HBO 2 and 4D-HBO 2 compared to the CTRL group ( Figure 5d). Other investigated proteins were similarly expressed among groups ( Figure 5).

HIF-1α mRNA and Protein Expression
Relative gene expression of HIF-1α transcription factor and its target gene VEGF was significantly increased in the 24H-HBO2 and 4D-HBO2 groups, compared to the CTRL and A-HBO2 groups ( Table 2), fold change (2 −ΔΔCT ) didn't showed significant changes ( Figure 4). Relative protein expression of HIF-1α was significantly increased in 24H-HBO2 and 4D-HBO2, compared to the CTRL group (Figure 5e).

HIF-1α mRNA and Protein Expression
Relative gene expression of HIF-1α transcription factor and its target gene VEGF was significantly increased in the 24H-HBO 2 and 4D-HBO 2 groups, compared to the CTRL and A-HBO 2 groups (Table 2), fold change (2 −∆∆CT ) didn't showed significant changes ( Figure 4). Relative protein expression of HIF-1α was significantly increased in 24H-HBO 2 and 4D-HBO 2 , compared to the CTRL group (Figure 5e).  Determination of vasoconstrictory 20-HETE showed its increased concentration in the A-HBO 2 group compared to CTRL and 4D-HBO 2 groups. There was a significantly decreased ratio of 14(15)-EET to 20-HETE, 11(12)-DiHET to 20-HETE and 8(9)-EET to 20-HETE in the A-HBO 2 group compared to other groups (CTRL, 24h-HBO 2 and 4D-HBO 2 ), and an increased ratio of 11(12)-DiHET to 20-HETE in the 4D-HBO 2 group versus the CTRL and A-HBO 2 groups (Table 4).

Discussion
The most important findings in the present study are: (a) a decrease in the contribution of NO to ACh-induced relaxation in HBO 2 -exposed animals; (b) improved hypoxia-stimulated relaxation after HBO 2 (24 h after single exposure and intermittent HBO 2 ); (c) for the first time, a demonstration that HBO 2 may increase production of 20-HETE and EETs in aortic tissue and increase the ratio of vasodilatory/vasoconstrictor metabolites (after intermittent HBO 2 ); (d) increased synthesis of HIF-1α and its involvement in hypoxia-stimulated relaxation of rat aortic rings, and (e) confirmation of EETs as endothelium-derived hyperpolarizing factors (EDHFs) and epoxygenases as oxygen sensors in HBO 2 .
Recently, our research group [7-9] showed the engagement of alternative pathways of endothelium-dependent relaxation in response to acetylcholine and ANG (1-7) in diabetic animals exposed to four days of HBO 2 . Interestingly, some previous studies have shown that if HBO 2 is applied more than once, adaptive mechanisms are activated to protect against further oxidative damage, e.g., increased antioxidative protection and general prevention against the genotoxic action of H 2 O 2 , mediated by increased protection by intracellular antioxidants of leukocytes and antioxidants, which scavenge ROS that are distant from nuclear DNA [23,24]. Consequently, HBO 2 preconditioning can be used for the prevention of subsequent oxidative injuries [25,26]. These HBO 2 -triggered adaptive and preconditional responses could also be responsible for changes in the underlying vascular reactivity mechanisms [7-9,27]. Since the effectiveness of HBO 2 treatments generally depends on exposures being repeated within a few days [28], it is especially important to define HBO 2 's molecular interactions when repeatedly applied.

Body Mass and Blood Pressure of Studied Groups
The observed decrease in mean arterial pressure is consistent with our previous studies, which discuss this issue [18,29]. Oxidative stress induced impaired baseline AChIR and HIR in the A-HBO 2 group, which was also previously described and discussed in our previous study [18].

The Mechanisms of Acetylcholine-Induced Vasorelaxation
The results show a decrease in the contribution of NO to ACh-induced relaxation in HBO 2 -exposed animals, despite the increased sensitivity of vascular smooth muscles to NO, which appears to be proportional to the duration of HBO 2 treatment (Figure 2b), as well as the equal contribution of COX-produced prostacyclin and CYP450 metabolites of AA in hypoxia-stimulated vascular relaxation ( Figure 3). As mentioned above, our previous findings, presented by Unfirer et al. and , also showed the presence of alternative pathways of endothelial relaxation to acetylcholine and ANG (1-7) in diabetic animals exposed to four days of HBO 2 , which represent the most likely avenues of increase in production or sensitivity to EETs. Most importantly, the discovery of the participation of CYP450 metabolites in vasorelaxation in HBO 2 is a novelty of the present study. These metabolites have been measured here for the first time ever in aortic tissue after HBO 2 treatment.

The Mechanisms of Hypoxia-Induced Vasorelaxation
A major part of the vasodilation response to hypoxia is mediated by the activation of COX and the consequent production of prostacyclin (PGI 2 ), which then activates K ATP channels [30]. Frisbee et al. [31] have shown that hypoxic vasodilation of skeletal muscular resistance arteries activates the cytochrome P450 enzyme pathway and that the hypoxic vasodilation response is not exclusively dependent on the COX pathway and prostacyclin production, but rather that EETs play a role in vasodilation in response to hypoxia. Some authors have shown that blood vessels with EDHF responses (coronary, cerebral and pulmonary) have a preserved EDHF response under severe hypoxia conditions, for which CYP450 epoxygenase is also responsible [32].
In the present study, the control group had an expected vasodilating response to hypoxia, whereas HIR was impaired after a single exposure to HBO 2 . Furthermore, 24 h after HBO 2 and after four days of repeated HBO 2 treatments (i.e., intermittent exposures), an increased vasodilating response to hypoxia appeared, suggesting that HBO 2 led to a change in the activation and/or expression of COX and/or CYP450 enzymes (Figure 3a). The vasodilation response of the control aorta was fully mediated by activation of COX and PGI 2 formation, because vasodilation was almost completely blocked after incubation with indomethacin. MS-PPOH reduced vasodilation at half the value of the baseline, indicating that CYP450 epoxygenase and the formation of EETs play a smaller role in vasodilation in hypoxia (Figure 3b). Twenty-four hours after single exposure to HBO 2 and after four consecutive days, incomplete inhibition of vasodilation occurs after the administration of indomethacin. There is an even greater inhibitory effect on vasodilation after the administration of MS-PPOH, indicating that HBO 2 induced higher EET production. Overall, these functional results suggest an important switch from the production of COX metabolites to CYP450 vasodilator metabolites after intermittent HBO 2 exposure.
The particular significance of present work is that we have measured eicosanoid production in aortic tissue for the first time ever. These measurements (Tables 3 and 4) support the results of functional studies with inhibitors. Acute HBO 2 exposure increased 20-HETE formation, which led to attenuated vasorelaxation in response to ACh and hypoxia. This explains the restored vasorelaxation in 4D-HBO 2 group and the altered mechanisms of vasorelaxation (with increased ratios of vasodilator to vasoconstrictor metabolites of CYP450 enzymes in this group), indicating that the vasodilation mechanisms are diverted to EETs. This is further supported by increased CYP2c11 gene and protein expression and an increased ratio of vasodilatory AA metabolites. Our results show increased concentrations of vasoconstrictor 20-HETE and a decreased ratio between various vasodilatory metabolites and 20-HETE in the A-HBO 2 group, but also show increased concentrations of vasodilatory metabolites in the A-HBO 2 and 24H-HBO 2 groups. Altogether, our results confirm epoxygenase and omega-hydroxylase as oxygen sensors, and reveal for the first time that EETs are the mediators responsible for recovery of vasorelaxation after intermittent HBO 2 treatment. Furthermore, the present study, with its direct measurements of CYP450 metabolites, confirms our hypothesis that periods between two hyperbaric exposures can be considered pseudohypoxia, which is a stimulus to enzyme upregulation and the alternation of vasoactive pathways [6,29]. The question of whether this is influenced by HIF-1α is discussed below.

The Role of HIF-1α in Vasorelaxation in Response to ACh and Hypoxia
Activation of HIF-1 involves redox-dependent stabilization of HIF-1α proteins [33]. In hypoxia, HIF-1α is translocated in the core and heterodimerized with a β subunit, forming the HIF-1 complex [34]. Two distinct domains within the HIF-1α subunit are responsive to cellular oxygen levels [35]. Expression and activation of the HIF-α subunits is firmly regulated, and their degradation by ubiquitin proteasome usually occurs in hyperoxic conditions [36,37]. However, independently of hypoxic or normoxic conditions, free radicals are needed to express HIF-α [36,37]. In acute HBO 2 exposure, the oxidative stress is increased [18]. Previously, it has been shown that HIF-1 and -2 levels are increased in HBO 2 treatment due to increased ROS production [37,38]. Using a DNA microarray, it has been reported that more than 2% of all human genes in arterial endothelial cells are regulated by HIF-1α, directly or indirectly [39], including genes for vascular reactivity and structural responses. For example, HIF-1α affects the expression of eNOS, iNOS, HO-1, COX-2, and the production of NO and prostaglandin under hypoxic conditions, such as hemorrhagic shock [40]. Importantly, changes in partial oxygen pressure modulate the synthesis of arachidonic acid metabolites [17,41]. The synthesis of epoxyeicosatrienoic acids (EETs), and also of 20-hydroxysacetoxyetraenic acids, is pO 2 -dependent and reaches its maximal plateau level at pO 2 from 80-150 mmHg. At lower values of pO 2 the formation of 20-HETE and EETs is reduced and is linearly dependent on pO 2 between 20 and 140 mm Hg, but the slope is less steep for EETs [17]. Whether HIF-1α has a role in changes of AA metabolite production in hyperbaric oxygenation [41] is still unclear, since there are contradictory results regarding HIF-1α expression in conditions of increased pO 2 , with various protocols of HBO 2 exposure [42][43][44][45]. However, in our functional experiments on aortic rings, we demonstrated the role of HIF-1α in the restored HIR. The blockade of HIF-1α expression attenuated HIR.
As already explained, the results of this study support our earlier hypothesis on the occurrence of pseudohypoxic conditions in pauses between two exposures. Although COX-1 and -2 gene expression decreased in the 24H-HBO 2 and 4D-HBO 2 groups, the protein expression of COX-2 ( Figure 5b) was increased, in accordance with the results of Kaidi et al. [46], who demonstrated that the upregulation of COX-2 is transcriptional and is associated with induction by HIF-1α. Results also indicate the association of HBO 2 , hypoxia, HIF-1α and increased EET production. Such results are consistent with the findings of Chen and Goldstein [47], suggesting a positive feedback mechanism that can explain induction and activation of CYP2C during hypoxia, although it remains unclear how EETs increase the expression of HIF-1α proteins and how phosphorylated AMPK activates transcription of the CYP2C gene. Suzuki and colleagues [39] have shown that the expression of mRNA HIF-1α does not increase with EETs and suggest that EETs stabilize HIF-1α by activating the PI3K/Akt path to induce the expression of VEGF. Taken together, the results of the present study suggest that improved relaxation in pseudohypoxic conditions after HBO 2 (after one exposure or four days of exposures) are in line with these studies. Functional results in the present study show the involvement of EETs in relaxation itself. It is plausible that HBO 2 , by increasing tissue pO 2 , induces oxidative stress. This oxidative stress consequently activates HIF-1α, as well as CYP2C11, which is followed by increased EET synthesis, which stabilizes HIF-1α. This further activates COX-2, to produce more prostacyclin and thus improve relaxation. Additionally, increased production of EETs or corresponding DiHETs serves as a substrate to increase COX activity. Once formed, epoxyeicosatrienoic acids (EETs) and 20-hydroxyeicosatetraenoic acid (HETE) are metabolized by β-oxidation to 18-and 16-carbon derivatives, which are less biologically active [48][49][50]. 5,6-EET, 8,9-EET and 20-HETE can also be metabolized by cyclooxygenase (COX) to vasoconstrictor endoperoxides or to vasodilator prostaglandin or prostacyclin-like derivatives [51]. Altogether, we have confirmed our hypothesis regarding the involvement of HIF-1α in HBO 2 effects. The importance of the present study is that results are highly conclusive in terms of the contribution of the metabolites of arachidonic acid (produced via CYP450 pathways) in the altered mechanisms of vasorelaxation. The present study demonstrates the importance of EETs and 20-HETE in vasorelaxation in healthy experimental animals, although we have also previously demonstrated their role in conditions such as stroke [52,53]. The present study lays the groundwork for the transition to human studies considering the mechanisms of the beneficial effects of HBO 2 in stroke.

Hyperbaric Oxygenation Exposure Protocols
The animals were bred and housed at the animal care facility of the Faculty of Medicine, Osijek. A total of 252 male Sprague-Dawley (SD) rats (aged 9-12 weeks) were used in this study. Rats were housed at a temperature of 21 • C-23 • C), in a humidity-and light-controlled room with free access to tap water and were fed ad libitum with a commercially prepared pellet diet (Mucedola, Italy). Hyperbaric groups underwent 120 min daily sessions of 100% O 2 at 2.0 bars absolute pressure at a flow rate of 2-3 L/min, with an additional 15 min for gradual compression and decompression in a Recompression Chamber for Experiments (110L, Djuro Djakovic, Aparati d.d., Slavonski Brod, Croatia). A-HBO 2 and 24H-HBO 2 groups underwent a single session and tissue sampling was performed immediately after decompression or 24 h after the single exposure, respectively. Animals from the 4D-HBO 2 group underwent one daily session for four consecutive days (intermittent) and tissue sampling was performed on the 5th day 24 h after the last exposure.

Measurement of Blood Pressure
Blood pressure measurement was performed using established protocols, as previously described [18,29,54,55]. Briefly, pressure values were measured using the Spacelabs Medical system (Spacelabs Medical, Inc., Redmond, WA, USA) using the PE-50 catheter inserted into the left femoral artery. The mean arterial pressure was calculated from the obtained values as a sum of systolic and double diastolic pressure and divided by three. As control values, the values measured in A-HBO 2 animals taken immediately prior to exposure to hyperbaric oxygenation were used. The same animals were used to collect thoracic aorta tissue for further analysis, including real-time quantitative PCR (RT qPCR) and/or protein expression by the Western blot method. Immediately after isolation, the aorta was transferred to a Petri dish with cold saline, cleansed from connective and fatty tissue and frozen in liquid nitrogen and then stored at −80 • C until analysis.

Experiments on Isolated Aortic Rings
The isolated aortic ring experiments were performed according to the well-established protocol in our laboratory [7-9,18]. After anaesthesia with ketamine 75 mg kg −1 and midazolam 0.5 mg kg −1 , thoracotomy was made and the thoracic aorta was isolated, cut to a 3-4 mm ring width and then placed in an organic pool (10 mL volume) with Krebs-Henseleit's solution (solution composition in mmol L −1 : 120 NaCl, 4.8 KCl, 1.2 KH 2 PO 4 , 2.5 CaCl 2 , 1.2 MgSO 4 , 25.5 NaHCO 3 , 10 glucose and 0.02 EDTA) continuously heated and oxygenated (t = 37 • C, pH = 7.4). After rinsing and stabilization for an hour, the endothelial preservation and maximum contraction tests were done, followed by protocols of acetylcholine-mediated relaxation or hypoxia-mediated protocols, with or without inhibitors.

Relative Gene Expression Determined by RT-qPCR Method
The relative expression of genes relevant to the studied mechanisms was determined by real-time quantitative PCR (RT-qtPCR; real time PCR, Bio Rad CFX96). Homogenization of the sample and total RNA was extracted with TRI reagent (Life Technologies, USA) according to the protocols of Chomczynski et al. [56]. RNA concentration and sample purity were determined using a Nanophotometer P300 UV/VIS, IMPLEN, and confirmation of the RNA presence was performed by placing samples on 1% agarose gel. Sample purification and cDNA preparation were performed according to the manufacturer's instructions from Sigma-Aldrich and Applied Biosystems.

Sample Preparation
Blood vessels were placed in to Precellys vials (2 mL), and 1 mL methanol with internal standard (IS) was added (concentration of IS was 50 ng mL −1 ). The samples were homogenized for 2 × 15 s at 4500 rpm and the next 20 s at 6000 rpm in a Precellys Evolution tissue homogenizer (Bertin Technologies, Montigny-le-Bretonneux, France). After homogenization, the samples were centrifuged for 5 min at 10,000 rpm in Thermo SL 40R centrifuge (Thermo Fisher Scientific, Waltham, MA, USA). The supernatants were moved to 1.4 mL polypropylene tubes and 200 µL of DMSO were added. The solvent was completely removed by means of a Biotage SPE DRY 96 sample evaporation system (Biotage AB, Uppsala, Sweden) [62].
Standards for calibration curve preparation: arachidonic acid (Sigma-Aldrich, from non-animal source, ≥98.5% (GC)) was dissolved in DMSO (10.0 mg mL −1 ) and diluted in methanol to 100 µg mL −1 . The Arachidonic Acid CYP450 Metabolite LC-MS Mixture (Cayman Chemicals Company, Ann Arbor, MI, USA) Arachidonic Acid CYP450 Metabolite LC-MS Mixture, Item No. 20665, declared to contain 10 µg mL −1 of each compound) was diluted in methanol to 100 µg mL −1 ( Table 5). The calibration range was from 0.001 to 50 ng mL −1 for all of standards except for the arachidonic acid, for which it was from 0.001 to 50 µg mL −1 . The calibration points were 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1, 5, 10 and 50 ng mL −1 (µg mL −1 for arachidonic acid, respectively). Quantification was performed by summing the three MRM signals of the same transition (differing for 0.001 Da) for each analyte in order to increase sensitivity. HPLC-electrospray ionization (ESI)/MS-MS: chromatographic separation was performed with a Waters Acquity UPLC ® BEH C18 column (2.1 × 50 mm; 1.7 µm particle size) (Waters, USA) and a gradient elution of mobile phase A (0.1% formic acid in water) and phase B (acetonitrile, at a flow rate of 400 µL min −1 ), according to the previously described method, modified as follows [63]. The acetonitrile part was 25% (v/v) at first, increasing to 50% within 7.5 min. The percentage of acetonitrile was continuously increased. At 12.5 min it was 60 percent and at 16 min it was 90 percent. It was held at 90% for a further 2 min. At 18.01 min the percentage of acetonitrile was reduced to 25 percent and was maintained at that level for 20 min. The column was reconditioned for 15 min with 75% of 0.1% formic acid in water and 25% acetonitrile ( Table 6). Column temperature was 50 • C and autosampler temperature was 4 • C. All measurements were carried out with a Shimadzu Nexera X2 HPLC system coupled to an AB Sciex QTRAP 6500 ion trap mass spectrometer from AB SCIEX (Framingham, MA, USA), operated under negative-ion electrospray (ESI) conditions in the SIM (Single (Selected) Ion Monitoring) and MRM (Multiple Reaction Monitoring) modes, respectively. Representative liquid chromatography-tandem mass spectrometry chromatograms are presented in Figure 6.

Statistical Analysis
All results are expressed as average ± standard deviation (SD). p < 0.05 was considered statistically significant. Acetylcholine-induced relaxation is expressed as a percentage of the maximum contraction. The response to ACh was analyzed via a two-way ANOVA with a post hoc Bonferoni test. For a comparison of mRNA and protein expression results and hypoxia-mediated relaxation of aortic rings, a one-way ANOVA variance test was used, or in the case of unequal distribution, Holm-Sidak or Kruskal-Wallis test data were obtained. For individual results, the determination of the difference between the normal distribution of numeric variables between the two independent groups was performed using Student's t-test, and in the case of a deviation from the normal distribution, the Mann-Whitney U test was used. SigmaPlot v.12 (Systat Software, Inc., Chicago, USA) and GraphPadPrism, Version 5.00 for Windows, GrafPad Software (San Diego, CA, USA) were used for statistical analysis.

Conclusions
In conclusion, intermittent HBO2 exposure causes enhanced relaxation, most likely by activating the CYP450 epoxygenases pathway of arachidonic acid metabolism and by increasing the formation of and sensitivity to EETs. In addition, HBO2 increases the expression of HIF-1α, which stimulates COX pathway expression and prostacyclin formation, thereby mediating enhanced relaxation in intermittent HBO2 exposure.  Acknowledgments: The authors are grateful to Sanja Novak, for help in animal handling and preparation. We thank Jasna Padovan, Director of Drug Metabolism and Pharmacokinetics Department, Fidelta Ltd., for support and valuable advice during the HPLC-ESI/MS-MS measurements. We also wish to thank Prof. Peter Nemeth, Lilla Prenek and Peter Engelmann, University of Pecs, Hu, for their help in performing western blot.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Statistical Analysis
All results are expressed as average ± standard deviation (SD). p < 0.05 was considered statistically significant. Acetylcholine-induced relaxation is expressed as a percentage of the maximum contraction. The response to ACh was analyzed via a two-way ANOVA with a post hoc Bonferoni test. For a comparison of mRNA and protein expression results and hypoxia-mediated relaxation of aortic rings, a one-way ANOVA variance test was used, or in the case of unequal distribution, Holm-Sidak or Kruskal-Wallis test data were obtained. For individual results, the determination of the difference between the normal distribution of numeric variables between the two independent groups was performed using Student's t-test, and in the case of a deviation from the normal distribution, the Mann-Whitney U test was used. SigmaPlot v.12 (Systat Software, Inc., Chicago, IL, USA) and GraphPadPrism, Version 5.00 for Windows, GrafPad Software (San Diego, CA, USA) were used for statistical analysis.

Conclusions
In conclusion, intermittent HBO 2 exposure causes enhanced relaxation, most likely by activating the CYP450 epoxygenases pathway of arachidonic acid metabolism and by increasing the formation of and sensitivity to EETs. In addition, HBO 2 increases the expression of HIF-1α, which stimulates COX pathway expression and prostacyclin formation, thereby mediating enhanced relaxation in intermittent HBO 2 exposure.