Hyperbaric oxygen therapy (HBOT) is a frequently applied remedy for soft tissue injuries [1
] and it seems to be an effective way of enhancing wound healing in diabetic patients [2
]. As a result of HBOT, plasma oxygen content and tissue oxygenation get into the normal range even in the depth of ulcerated wounds. HBOT also mobilizes stem cells, such as endothelial progenitor cells, therefore, may represent a therapeutic aid for revascularization [3
]. However, as available oxygen content increases in the core of the body, cardiovascular oxidative damage may emerge as a possible consequence of HBOT [4
Elevated oxidative-nitrative stress can be detected in diabetic patients and alternative glucose metabolic pathways emerge. As a result, inflammatory processes are initiated and a further increment of oxidative-nitrative stress is detectable too. All these processes contribute to increasing the probability of cardiovascular damage. Therefore the possible adverse effects of hyperbaric oxygen therapy (HBOT) on oxidative balance and consequently on the cardiovascular system in diabetes raise questions [5
The effects of hyperoxia on cardiovascular parameters and cytokine production are contradicting. Elevation of reactive species formation during HBOT is proven in animal models and human studies [6
] but it has not been established yet whether this elevation is preserved days or even weeks after the cessation of HBOT. The concern is that elevated reactive radical formation may lead to the worsening of chronic inflammation and cardiovascular status. According to the metanalysis of de Smet et al., the short-term outcome of HBOT is a significant reduction of wound area; however, the research team warns us about the possible long-term side effects due to increased oxidative stress [9
]. On the other hand, HBOT has a hormetic effect that may lead to increased antioxidant capacity. It is sought that HBOT reduces neutrophil-endothelial adhesion [10
] through inhibiting neutrophil recruitment; also, matrix metalloproteinase activation is abated [11
]. Therefore, it is possible that the net result of HBOT is decreased inflammation and consequently improved cardiovascular function [12
Poly(ADP-ribosyl)ation (PARylation) is a ubiquitous consequence of emerging oxidative and nitrative stress. Superoxide reacts with nitric oxide (NO), resulting in peroxynitrite (ONOO−
) formation and the produced reactive species attack DNA, causing single- and double-stranded breaks. These breaks are recognized by poly(ADP-ribose) polymerase 1 (PARP), that builds up long, branching chains of poly(ADP-ribose) (PAR) from NAD+
. The subsequent loss of NAD+
leads to ATP depletion and cell death. Lower but steady PARylation of nuclear proteins may help DNA repair but also increase NF-κB expression and cytokine production; therefore constant PARP activation leads to chronic inflammation too. Increased cytokine production activates immune cells and keeps up oxidative and nitrative stress, leading to a vicious cycle [13
Previously it was shown that PARylation increased following hyperbaric oxygen treatment [15
] only when 50 atm of oxygen pressure was applied. In a rat model of severe acute pancreatitis, HBOT along with PARP-inhibition had an additive effect [16
]; however, long-term consequences of HBOT on PARylation has not been assessed before.
In the presented study, we estimated the effect of HBOT in a rat model of fully developed type1 diabetes on the cardiovascular system, cytokine production, oxidative stress and subsequent PARylation.
2. Materials and Methods
All investigations conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 85–23, Revised 1985) and was approved by Veterinary and Food Administration Institute of Budapest (1896/003/204).
A total of forty male Wistar rats weighing around 300 g were used for the experiment. They had water and standard rat chow ad libitum during the course of the experiment. They were randomly distributed into four groups, 10 rats per group, as follows—no diabetes and no HBO treatment (Control); HBO treatment without diabetes (Control HBOT); Diabetes without HBO treatment (DM); and finally diabetes with HBO treatment as well (DM HBOT). Up to two animals per group deceased between the 3rd and 6th weeks of the experiment.
Twenty rats (future DM and DM HBOT) were injected intravenously with 70 mg/kg streptozotocin dissolved in citrate buffer (both Sigma Aldrich, St. Louis, MO, USA), under thiopentone sodium (Euthasol, Phylaxia-Sanofi, Hungary) anesthesia. Three weeks after the initiation of diabetes 24-hour blood glucose profile was taken using capillary blood samples obtained from the tail in every six hours from 8 AM. Ten control and 10 diabetic rats were enrolled in a daily hyperbaric oxygen treatment (Control HBOT) regimen 3 weeks after diabetes induction. HBO exposure was set at a pressure of 2.5 Bar for 60 min (the appropriate pressure was attained in 5 min, depressurization took an additional 8 min, this 13-minute period is not included in the 60-minute treatment period) on Thursday and Friday on the first week and the second and third week from Monday to Friday (12 days in the course of 16 days, no treatment on weekends; as it was commonly applied in human patients at Baromedical Ltd. (Budapest, Hungary; a Hungarian medical company providing HBOT) to decrease the risk of pulmonary side effects, Figure 1
The day after the completion of the full HBO treatment protocol twenty-four-hour blood glucose profile of the rats was taken again.
To assess the long-term effects of HBOT, two weeks after the completion of the HBOT regimen 24-hour blood glucose profile and HbA1c levels were measured and the animals underwent echocardiographic measurement under halothane anesthesia (both M-mode and two-dimensional mode, 7–15 MHz, Hewlett Packard Sonos 5500, San José, CA, USA), end-diastolic and stroke volume (EDV and SV), ejection fraction (EF) and fractional shortening (FS) were calculated. Afterward, we sacrificed the rats in order to perform ex vivo measurements and sample collection (blood, whole hearts and thoracic aortae). The thoracic aorta was cleared from the surrounding periadventitial fat and was cut into 3–4 mm wide rings, laid in organ baths that was filled with warmed (37 °C) and oxygenated (95% O2, 5% CO2) Krebs’ solution (CaCl2 1.6 mM; MgSO4 1.17 mM; NaCl 130 mM; NaHCO3 14.9 mM; KCl 4.7 mM; KH2PO4 1.18 mM; Glucose 11 mM). Isometric tension was measured using isometric transducers (DMT, Hinnerup, Denmark) and digitized, stored and displayed (Biopac, Goleta, CA, USA) on a personal computer. A basal tension of 15 mN was applied and the rings were equilibrated for 60 min, vascular contractility was determined by phenylephrine dose-response curves (Phe, 10−9 to 3 × 10−4 M). The rings were allowed to equilibrate and to restore basal tone for 60 min. Afterward, phenylephrine precontraction (10−6 M) was induced and relaxation ability was determined by acetylcholine dose-response curve (Ach, 10−9 to 3 × 10−4 M) after). From each experimental group, 5 to 6 pairs of rings were gained and used during this experiment.
Heparinized and clotted whole blood samples were used for plasma and serum collection, respectively. The heparinized blood samples were gradient centrifugated on Histopaque-1083 (Sigma Aldrich, St. Louis, MO, USA) to isolate mononuclear blood cells, which were smeared on frosted glass microscopic slide and fixed in methanol. Intact aortic segments and hearts were fixed in 4% formaldehyde solution and then paraffin-embedded sections were cut.
After deparaffinization of the sections, antigen retrieval (80 °C for 15 min in 0.1 M citrate buffer, pH 3) and blocking of endogenous peroxidase activity, the samples were incubated overnight at 4 °C with monoclonal anti poly(ADP-ribose) antibody (PAR, made in mouse, Tulip Biolabs, West Point, PA, USA, 1:500) or monoclonal anti poly (ADP-ribose) polymerase (PARP) antibody (Cell Signaling Technology, Danvers, MA, USA, 1:100). Secondary labeling was achieved using a biotinylated anti-mouse horse antibody (Vector Laboratories, Burlingame, CA, USA) (30 min, room temperature). Horseradish peroxidase-conjugated avidin (ABC kit, 30 min, room temperature, Vector Laboratories, Burlingame, CA, USA) and nickel-enhanced diaminobenzidine (6 min, room temperature, Vector Laboratories, Burlingame, CA, USA) were used. Tissue sections were counterstained with nuclear fast red for PAR (Reanal, Budapest, Hungary) or hematoxylin (PARP). To assess staining intensity, we captured 5 microscopic fields (at 200-fold magnification) of each sample and the percentage of the dye-positive area was determined in either ventricular wall area (cardiac samples) or endothelial cell layer area (aortae) by a blinded experimenter. Image analysis was done by ImageJ (1.49v NIH, Bethesda, MD, USA), the background of the original photos was subtracted, if present, debris was erased only from the background of the samples and a 2-bit conversion was applied. The threshold was set to the same value in case of every photo to measure positive and total area. No other manipulation of the pictures was done after taking the micrographs. The presented representative sample micrographs were taken at 400-fold magnification and the composite of the eight pictures was manipulated by applying a (common) histogram adjustment layer to achieve an evenly white background and acceptable contrast.
For Western blot analysis, cardiac tissue was ground in liquid nitrogen with a mortar and pestle and suspended in ice-cold lysing solution (30 mM Na-HEPES, 100 mM NaCl, 2% (w/v) Triton-X-100, 20 mM NaF, 1mM Na-EGTA, 1 mM Na-EDTA, 100 mM benzamidine, 0.02% (w/v) diisopropyl phosphorofluoridate, 1% (w/v) aprotinine, 1% (w/v) protease inhibitor cocktail (Sigma-Aldrich), 1% (w/v) phosphatase inhibitor cocktail 2 (Sigma-Aldrich, San José, MO, USA) and 1% (w/v) phenyl methyl sulfonyl fluoride; pH 7.5) at a ratio of 250 μL of buffer per 0.5 g tissue (wet mass). Equal amounts of protein (50 μg) per lane were subjected to 10% (w/v) SDS-polyacrylamide gels and transferred to nitrocellulose membranes (GE Healthcare, Chicago, IL, USA) at 100 mA/cm2 for 2 h. After blocking with 5% (w/v) blotting grade (Bio-Rad Laboratories, Hercules, CA, USA) dissolved in PBS, supplemented with 0.1% (w/v) Tween 20, membranes were decorated with monoclonal anti poly (ADP-ribose) polymerase (PARP) antibody (Cell Signaling Technology, Danvers, MA, USA) overnight at 4 °C in 1:1000 dilution. Bound antibody was detected with enhanced chemiluminescence using horseradish peroxidase-conjugated anti-rabbit-Ig (from donkey) secondary antibody (GE Healthcare, Chicago, IL, USA) used in 1:5000 dilution. ImageJ software (NIH, Bethesda, MD, USA) was used for densitometry analysis. PARP was normalized against total protein performed by densitometry of Ponceau S stained membranes.
Systemic oxidative stress was gauged by malonyl-dialdehyde assay. MDA content of the plasma was detected based on the thiobarbituric acid reactive substances (TBARS) assay. The MDA-TBA product formed by the reaction of MDA and thiobarbituric acid (TBA) under high temperature (90–100 °C) and acidic conditions was measured at 540 nm [8
To identify the cytokines involved rat cytokine array was performed (R&D Systems, Minneapolis, MN, USA) on plasma samples of a representative rat from each group (based on blood glucose levels, heart functions and vascular reactivity). After visual evaluation with the naked eye, the most affected cytokines-cytokine-induced neutrophil chemoattractant-1 (CINC-1), tissue inhibitor of metalloproteinases-1 (TIMP-1) and lipopolysaccharide-inducible CXC chemokine (LIX) were measured from each collected sample using commercially available ELISA kits according to the users’ manual (R&D Systems, Minneapolis, MN, USA).
Results are reported as mean ± SEM. To investigate Gaussian distribution, D’Agostino and Pearson’s omnibus normality test was performed. In case of non-Gaussian distribution, logarithmic transformations were performed. This transformation resulted in normal distribution for CINC-1 and LIX levels. Statistical significance between two measurements was determined by the two-tailed unpaired Student’s t-test (HbA1C) and among groups, it was determined by two-way analysis of variance (ANOVA) with Tukey’s post hoc test. In case of differing variances between groups (area under glucose curve, MDA, CINC-1 and TIMP), nonparametric test (Kruskal-Wallis) was also applied with Dunn’s post hoc test. Probability values of p ≤ 0.05 were considered significant.
Our aim was to assess the long-term consequences of hyperbaric oxygen treatment on cardiovascular and redox status of type 1 diabetic rats. In our study, we used a widely accepted animal model. Cardiac parameters of the diabetic rats were significantly worse than those of the controls, which suggests the presence of diabetic cardiomyopathy. Moreover, the deterioration of NO-mediated relaxing ability of aortae indicates that endothelial dysfunction also developed. Oxidative status, PAR-ylation and inflammatory signals were also elevated compared to control animals. These manifestations of diabetic symptoms are well-known and counted toward the major contributors of diabetic complications.
Carbohydrate household and cardiovascular status of control rats undergoing hyperbaric sessions were not affected by the hyperoxia and their TBARS and inflammatory cytokine production did not show any alteration either. This suggests that non-diabetic patients are safe in this regard under the hyperoxic environment. Furthermore, it is sought that the development of type 1. diabetes mellitus can be postponed by HBOT, as it was shown in an animal model of autoimmune diabetes.
The study revealed that HBOT did not increase, however slightly decreased the blood glucose levels of diabetic animals, shown by the reduction of AUC glucose. Many previous studies found that HBOT leads to an elevation of blood glucose levels of streptozotocin-induced diabetic rats. First, it was thought that this elevation that follows augmented oxidative stress is a direct consequence of hyperoxia in a diabetic system [17
]. For instance, Matsunami et al. described an elevation of blood glucose level parallel with severe alterations in the expression of antioxidant enzymes under HBOT in a rat model of streptozotocin-induced diabetes, concluding that HBOT further damages carbohydrate household and increases oxidative stress [18
]. Later it was recognized that streptozotocin treatment itself causes a vast elevation of reactive oxygen species as part of its mechanism of action [21
]. Therefore, one may speculate that streptozotocin-induced and naturally developing diabetes may react differently to HBOT. This hypothesis is supported by Faleo et al., who described a slower development of type 1. diabetes in non-obese autoimmune diabetic mice [22
]. To prevent the confounding effect of streptozotocin treatment, our team decided to start hyperbaric sessions after the complete degradation/excretion of streptozotocin. Interestingly, other observations of type 2 human subjects suggested that carbohydrate household may be improved by HBOT [23
]. Accordingly, we also observed decreased AUC glucose and a tendency of decrement in HbA1c in our model but—possibly due to the short timeframe of our study—it did not reach the level of significance.
Our study also showed that HBOT did not deteriorate the cardiovascular status of diabetic animals; however, oxidative stress, poly(ADP-ribozyl)ation and inflammatory processes were dulled.
Subclinical inflammation is present in diabetic patients. One typical executor of the chronic inflammation is PARP activity leading to elevated inflammatory cytokine (e.g., TNF alpha) production and consequential NF-κB expression. Poly(ADP-ribose) polymerase 1 (PARP) is activated as a response to DNA breakage in elevated oxidative stress. This increment of inflammatory transmitters leads to a further increment of free radical formation. Also, PARP plays a role in arresting or initiating cell death, promoting repair mechanisms and cooperating with transcription factor formation (reviewed by References [24
]. As poly(ADP-ribosyl)ation was only elevated in untreated diabetic rats, our results suggest that hyperoxia indirectly inhibits PARP activation and the consequent inflammatory vicious circle. As the muted TBARS product suggests, we can hypothesize that the link between lower PAR-ylation and hyperbaric treatment may possibly be the activation of antioxidant mechanisms, as similar protection was achieved by Ayvaz et al. in a bile duct ligation model [26
]. The suggestion that hyperbaric oxygen treatment and inhibited poly(ADP-ribosyl)ation cooperate in conserving tissue integrity comes from Inal et al. [16
] who proved that oxidative stress parameters and histopathology are ameliorated by PARP inhibition and HBOT but their effects are increased when acting together in severe acute pancreatitis. In their model, acute pancreatitis caused a loss of superoxide-dismutase and glutathione peroxidase along with an elevation in MDA levels. Induction of pancreatitis was followed by two sessions of HBOT and data were collected after two sessions and two days. Interestingly, according to their findings HBOT acutely decreased oxidative stress, increased antioxidant capacity and protected tissue functions. According to our results, this theory can be applied to diabetes as well. In addition, inhibition of PARP resulted in a similar protection in acute pancreatitis; therefore, it was suggested that diabetic rats benefit from PARP inhibitors (e.g., References [27
Our group assessed endothelial function in the thoracic aortae. We accepted the assumption that peripheral vessels, mainly microcirculatory units, are similarly affected by hyperglycemia. One possible explanation to the worsening of endothelial function of hyperglycemic patients is the uncontrolled glucose uptake of endothelial cells and consequential mitochondrial overproduction of oxidative agents [30
]. The role of skeletal muscle oxygenation in metabolic control is also influential. Ameliorated skeletal muscle oxygen supply described by Yamakoshi et al. [31
] can be one reason behind improved inflammatory status and metabolic changes. In comparison, effects of training may have similar influence than that of HBOT on cardiovascular and inflammatory status of diabetic patients, because physical activity also elevates immediate oxidative stress and activates antioxidant mechanisms [32
Increased activity of matrix metalloproteinases (MMPs), along with elevated oxidative stress, is proposed to be one of the leading causes behind the higher risk of ischemic events in diabetic patients. In a HBOT preconditioning model, it was shown that hyperbaric environment leads to a decrement of MMP-2 and 9 in an animal stroke model [33
]. Also, previously, HBOT was proven to inhibit NF-κB expression in healthy rats. Due to the low number of samples and high variances of CINC-1 and TIMP-1 levels in the diabetic animals, our results do not allow to reinforce these observations; nonetheless, the presented findings do not point toward an increment in inflammatory processes after HBOT. Further inquiries may reveal a cohort, which could benefit HBOT as an anti-inflammatory intervention.
Cardiac dysfunction developing in maltreated diabetes can be the result of insufficient oxygen supply by the coronary circulation, subclinical systemic inflammation and elevated oxidative-nitrative stress and an increment of alternative metabolic pathway byproducts (recently reviewed by Reference [34
]). Diastolic dysfunction develops earlier than systolic dysfunction. As hyperbaric treatment ameliorated the cardiac oxygen supply and alleviated inflammatory cytokine production, this could have been one reason of maintained diastolic filling. Regulation of cardiac function better maintained in HBOT animals due to the slower deterioration of autonomic functions [31
Recently, it was shown that HBOT increased oxygenation in injured muscle tissue up to 24 hours after hyperbaric exposure, along with a reduction in edema formation. Better oxygenation of the affected tissues may contribute to the healing of the ulcerated wound. Furthermore, activation of the pro-inflammatory IL-6/STAT3 pathway is sought to participate in the inflammation following soft tissue injury. The peak of IL-6 production was detected prior in HBOT animals than in non-treated rats leading to earlier STAT3 phosphorylation. According to the team of Yagishita, STAT3 phosphorylation is a major step in satellite cell activation and tissue repair. Whether in diabetic patients, the progenitor cell activation contributes toward wound healing and conservation of cardiovascular status, needs to be further assessed [35