Secondary Logo

Journal Logo


Moderate hyperoxia induces inflammation, apoptosis and necrosis in human umbilical vein endothelial cells

An in-vitro study

Hafner, Christina; Wu, Jing; Soto-Gonzalez, Lourdes; Kaun, Christoph; Stojkovic, Stefan; Wojta, Johann; Tretter, Verena; Markstaller, Klaus; Klein, Klaus U.

Author Information
European Journal of Anaesthesiology: March 2017 - Volume 34 - Issue 3 - p 141-149
doi: 10.1097/EJA.0000000000000593
  • Free



Supplementary oxygen (O2) is given routinely in the perioperative setting to prevent hypoxia in vital organs. Excessive O2 may cause systemic hyperoxia that can give rise to damage of pulmonary, cardiovascular, immune and nervous tissues through a disturbance of cellular redox homeostasis.1 Recent studies in patients undergoing general surgery suggest that hyperoxic exposure is associated with higher mortality and morbidity.2–4 A recent post-hoc analysis of the PROXI trial suggests that perioperative hyperoxia may be associated with complications that include myocardial infarction (MI) and acute coronary syndrome.5 Also, a growing body of evidence suggests that arterial hyperoxia may be linked to increased in-hospital mortality in a dose-dependent fashion.6,7 Accordingly, current guidelines of the European Resuscitation Council recommend that hyperoxia is avoided and that O2 therapy is provided according to individual need after the return of spontaneous circulation.8,9

The risks of a hyperoxic episode during O2 therapy are elevated in surgical and intensive care patients with normal cardiac and pulmonary function. At the same time, cyclic hyperoxia/anoxia may result from recruitment and derecruitment of lung atelectasis (cyclic atelectasis) within the respiratory cycle during higher dose O2 administration.10–13 Both prolonged constant hyperoxia and cyclic hyperoxia/anoxia may harm the vascular endothelium. In human umbilical vein endothelial cells (HUVECs), extreme constant O2 (95%) or cyclic hyperoxia/hypoxia (95%/5%) can induce cellular inflammation, apoptosis and necrosis.14 Given the vast surface area of the human vascular endothelium, approximately 3000 to 5000 m2, systemic hyperoxia may be an important contributor to the adverse cardiovascular outcome recently reported in patients receiving therapeutic higher dose O2.4,5 At present, however, it remains uncertain if moderate, as opposed to extreme constant or cyclic hyperoxia, can also injure the human vascular endothelium.

The current study aims to investigate the effects of moderate constant (40% O2) and cyclic hyperoxia/anoxia (40%/0% O2) on HUVEC growth, viability and inflammatory responses compared with those of control normoxia (21% O2) and constant anoxia (0% O2). We speculated that moderate hyperoxia inhibits cell proliferation and induces proinflammatory cytokine release and cell death. We also conducted a kinase array assay to identify potential cellular signalling mechanisms activated by different O2 regimens.


Cell culture

Ethical approval (Ethical Committee No. 1616/2013) was provided by the Medical University of Vienna Ethics Committee, Vienna, Austria on 23 September 2013. After written consent, HUVECs were isolated from fresh umbilical cords by mild collagenase treatment as described previously.14,15 HUVECs were chosen for this series of experiments because they are an accepted model for the vascular endothelium.16 Cells were cultivated in M199 medium (Gibco, Invitrogen; Carlsbad, California, USA) together with 20% foetal calf serum (FBS superior, Biochrom, Berlin, Germany), 150 μg ml−1 bovine pituitary endothelial cell growth supplement (ECGS, Sigma-Aldrich; St Louis, Missouri, USA), 5 U ml−1 heparin and 1% penicillin/streptomycin (Gibco, Invitrogen). Twenty-four hours before the start of treatment, HUVECs were seeded onto gelatin (Sigma-Aldrich) pre-coated cell culture plates with respiratory active membranes (OptiCell plates, Thermo Fisher Scientific; Waltham, Massachusetts, USA) at 5 × 105 cells per plate.

Study design

The OptiCell plates consist of two 75-μm gas-permeable respiratory active membranes that allow cell growth on the inner side of the membranes surrounded by medium. Ten plates were placed in a custom-made framed box connected to a gas distributor and pre-mixed bottles containing normoxic gas (21% O2/5% CO2/74% N2), high O2 gas (40% O2/5% CO2/55% N2) and anoxic gas (0% O2/5% CO2/95% N2) to create normoxic, constant hyperoxic, cyclic hyperoxic/anoxic and constant anoxic O2 exposure conditions, respectively.17 For cyclic hyperoxia/anoxia, the hyperoxic and anoxic gas mixtures were connected to an electronic gas flow meter and a computer-controlled valve (EL-Flow Select Series; Bronkhorst, Netherlands) to switch between 40 and 0% O2. The O2 oscillation (hyperoxia/anoxia) was set to three complete cycles per hour (1 cycle = 10 min 40% O2 then 10 min 0% O2). This frequency was chosen to make sure that the cells experience the full amplitude of oscillations at maximal frequency.17 All bioreactor units, their gas supply tubes and gas humidifiers were kept in a standard incubator (Heratherm, Thermo Fisher Scientific at 37 °C. Temperature and relative humidity in each unit were controlled using an electronic sensor (SHT-71, Sensirion; Zürich, Switzerland).

Cell growth and viability

OptiCell plates were removed from the bioreactor after 0 (baseline), 6, 12, 24 and 48 h (Fig. 1). Cell culture supernatants were collected, and lactic dehydrogenase release was measured using a colorimetric enzymatic assay (Cytotoxicity Detection Kit, Roche; Mannheim, Germany). Optical density was measured at 490 nm with a spectrophotometer (Viktor 3, PerkinElmer, Waltham, Massachusetts, USA). Cells were harvested, centrifuged at 200 × g and re-suspended in 500 μl phosphate buffer solution. Cells were counted using a haemocytometer (Hausser Scientific; Horsham, Pennsylvania, USA) after trypan blue staining (0.4% trypan blue, Sigma-Aldrich). Cell suspensions were stained with annexin V and propidium iodide using an apoptosis detection kit (BD Pharmingen; Heidelberg, Germany) to identify necrotic and apoptotic cells by fluorescence-activated cell sorting (FACS) (Beckman Coulter).

Fig. 1
Fig. 1:
Experimental protocol.

Cell inflammation and injury

The inflammatory response was measured by ELISAs for IL-6, IL-8 and macrophage migration inhibitory factor (MIF) (DuoSet, R&D Systems; Minneapolis, Minnesota, USA). A phosphokinase array was performed after exposure to different O2 concentrations (R&D Systems).

Statistical analysis

All experiments were performed three times using separate cultures with three internal repeats. Data are expressed as mean ± SD. Analyses and plots were made using Prism 6.0 software (GraphPad; San Diego, California, USA). Two-way analysis of variance was used to compare treatment group means at various time points, followed by Bonferroni's post-hoc analysis for multiple comparisons. A P value less than 0.05 was considered significant.


The current in-vitro study was not registered.

Cell culture

Stable O2 conditions were confirmed for constant normoxia (21 ± 0.5% O2), constant anoxia (0 ± 0.5% O2) and constant hyperoxia (40 ± 0.5% O2) units. Oscillating O2 was produced in the cyclic hyperoxia/anoxia unit (40%/0% O2, average 20% O2) at three full oscillation cycles per hour. Temperature (37.0 ± 0.5 °C), relative humidity (95 ± 5%) and ambient pressure (<0.1 kPa increase due to constant or cyclic gas flows) remained stable during experiments.

Cell growth and viability

Phase contrast microscopy revealed that cultures of HUVECs exposed to constant hyperoxia, cyclic hyperoxia/anoxia or constant anoxia for 48 h exhibited fewer polymorphic healthy cells and greater numbers of round phase-bright injured cells than cultures incubated under normoxia (Fig. 2). Cell growth and viability were assessed quantitatively by trypan blue exclusion and FACS analysis of annexin V/propidium iodide staining (Fig. 3). There was a significant reduction in viable cell count after 12 and 24 h exposure to constant hyperoxia compared with exposure to constant normoxia (P < 0.05). Likewise, there was a significant reduction in cell count after 48 h exposure to cyclic hyperoxia/anoxia and constant anoxia (P < 0.05). Trypan blue staining showed larger proportions of dead cells under constant hyperoxia (P < 0.05) and cyclic hyperoxia/anoxia (P < 0.01) than under normoxia. FACS analysis of annexin V/propidium iodide staining revealed a decrease in live cells (P < 0.01) and concomitant increases in apoptotic cells (P < 0.01) and necrotic cells (P < 0.01) after 48 h constant hyperoxia compared with constant normoxia. Cyclic hyperoxia/anoxia also induced a significant reduction in live cell count and increases in both trypan blue–stained cells and necrotic cells at 48 h (P < 0.01) compared with normoxic controls, although the effects were generally milder than those observed after constant hyperoxia. In contrast, constant anoxia appeared to reduce the cell count after 48 h (P < 0.01), indicative of an effect on proliferation, but did not increase the numbers of apoptotic cells.

Fig. 2
Fig. 2:
Morphological signs of human umbilical vein endothelial cell degeneration under moderate hyperoxia and anoxia. (a) to (d) Changes in cell morphology after 48 h exposure to different oxygen concentrations. Human umbilical vein endothelial cells exposed to constant moderate hyperoxia (40%, c) or cyclic hyperoxia/anoxia (40%/0%, d) exhibited fewer polymorphic viable cells and more numerous injured (rounded and phase-bright) cells than cultures exposed to constant anoxia (b) or normoxia (a).
Fig. 3
Fig. 3:
Growth inhibition and induction of multiple cell death pathways by moderate hyperoxia. Panels (a) to (e): cell growth, trypan blue–positive cells and fluorescence-activated cell sorting analysis of live, apoptotic and necrotic cells after exposure to different oxygen concentrations. * P < 0.05, versus constant normoxia (21% oxygen) at indicated time point; + P < 0.05, versus constant anoxia (0% oxygen) at indicated time point; # P < 0.05, versus constant hyperoxia (40% oxygen) at indicated time point; $ P < 0.05, versus all other conditions at indicated time point.

Release of inflammatory cytokines

Cytokine secretion patterns are displayed in Fig. 4. Cells exposed to 48 h of constant hyperoxia showed markedly increased release of IL-6, IL-8 and MIF compared with cells under constant normoxia (all P < 0.05). Cells exposed to 48 h of cyclic hyperoxia/anoxia showed increased IL-6 and IL-8 release (all P < 0.05).

Fig. 4
Fig. 4:
Induction of proinflammatory cytokine release by moderate hyperoxia. Panels (a) to (c): release of IL-6, IL-8 and macrophage migration inhibitory factor after exposure to different oxygen concentrations. * P < 0.05, versus constant normoxia (21% oxygen) at indicated time point; + P < 0.05, versus constant anoxia (0% oxygen) at indicated time point; § P < 0.05, versus all other conditions at indicated time point.

Differential activation of kinase pathways

Phosphokinase array results are displayed in Fig. 5. Each O2 exposure regimen induced a distinct pattern of kinase activation. Constant hypoxia markedly increased the phosphorylation/activation of mitogen-activated protein kinase (MAPK) kinases p38a, extracellular regulated kinase 1/2, c-jun kinase 1/2/3 and the mitogen-activated and stress-activated kinase compared with all other treatments (all P < 0.05). Constant hypoxia also activated Lyn, Lck, Fgr and p70S6, signal transducers and activators of transcription (STATs) 2, 5b and 6, glycogen synthase kinase (GSK) 3a/b and 5′ AMP-activated protein kinases (AMPKs) 1/2 compared with constant normoxia (all P < 0.05). Constant hyperoxia exposure increased p38a, Lyn and p70S6, whereas cyclic hyperoxia/hypoxia exposure increased p38a, Lyn, p53, Akt, mechanistic target of rapamycin (TOR) complex 1, p70S6, PRAS40, STAT6, GSK 3a/b and AMPKa1/2 compared with normoxia.

Fig. 5
Fig. 5:
Activation of multiple cytoprotective and cell death–associated kinase pathways by moderate hyperoxia. Panels (a) to (g): phosphokinase array showing changes in mitogen-activated protein kinases, Src, p53, heat shock protein, Akt; target of rapamycin, p70, PRAS, signal transducer and activator of transcription isoforms, glycogen synthetase kinase and 5′ adenosine monophosphate–activated protein kinase activation after 60 min exposure to different oxygen concentrations. * P < 0.05, versus constant normoxia (21% oxygen) at indicated time point; + P < 0.05, versus constant anoxia (0% oxygen) at indicated time point; # P < 0.05, versus constant hyperoxia (40% oxygen) at indicated time point; § P < 0.05, versus all other conditions at indicated time point.


Key findings

HUVECs subjected to constant mild and cyclic hyperoxia/anoxia for longer than 6 to 12 h gradually developed signs of injury, shown by decreased cell count and increased numbers of trypan blue–positive, apoptotic and necrotic cells. After longer than 24 to 48 h of treatment under all three conditions investigated proinflammatory cytokines were released. Phosphokinase array analysis showed that 60 min of constant moderate hyperoxia, cyclic hyperoxia/anoxia or constant anoxia activated distinct kinase pathways, which may account for the differences in cell death. Our findings suggest that the vascular endothelium may be injured by moderate constant and cyclic hyperoxia, but also normoxic exposure showed some deleterious effects on HUVECs. For these reasons, the findings of the present in-vitro study must be interpreted with caution and cannot be transferred to a clinical scenario.

Sensitivity of the vascular endothelium to hyperoxia

The vascular endothelium lines the interior surface of all blood and lymphatic vessels and forms the interface between blood and tissue. It is critical for maintaining tissue metabolic homeostasis. Endothelial functions include fluid filtration, regulation of blood vessel tone, haemostasis, neutrophil recruitment, hormone trafficking and angiogenesis.18 Numerous factors associated with surgery may cause endothelial dysfunction, including tissue trauma and altered blood O2.19,20 A recent in-vitro study showed that extreme constant hyperoxia (95% O2) or cyclic hyperoxia/hypoxia (95%/5% O2) gradually induced inflammation, apoptosis and necrosis of HUVECs.13

Mechanisms by which hyperoxia causes injury to the vascular endothelium include reactive O2 species (ROS)-dependent signal transduction, leading to endothelial activation and inflammation with ensuing barrier disruption and cardiovascular dysfunction. Given that the surface of the vascular endothelium (3000 to 5000 m2) is approximately 50 times that of the pulmonary endothelium (70 to 90 m2), deleterious effects on the endothelium may have an important pathogenic influence during the perioperative phase.21 However, such extreme hyperoxic ranges (e.g. 5 to 95%) do not occur in clinical settings. It remains unknown if moderate hyperoxia can induce endothelial dysfunction, but we did observe a significant increase in the proportions of apoptotic and necrotic HUVECs after 48 h of constant moderate hyperoxia, shown by a progressive increase in trypan blue–positive (nonviable) cells. Also, the proportion of trypan blue–positive cells was increased when compared with that of the normoxic controls after only 6 h. Cyclic hyperoxia/anoxia also induced significant injury to HUVECs, as shown by reduced cell numbers, decreased live cells and increased necrotic and trypan blue–positive cells after 48 h. Rates of cell death were generally higher than in cultures exposed to anoxia. This finding suggests that endothelial cells might be more prone to acute and chronic hyperoxic injury than to hypoxic injury.22 Normoxic exposure also impaired cell viability in HUVECs, something we can most probably attribute to our cell culture model. We present a novel hypothesis of injury, hence results should be interpreted with caution and need confirmation.

Inflammation and injury of the vascular endothelium under hyperoxia

Our results also show that release of the proinflammatory cytokines IL-6 and IL-8 is substantially increased during constant moderate hyperoxia and cyclic hyperoxia/anoxia, though they appear later than the decrease in vitality. The production and activity of these cytokines represent an important pathway in endothelial inflammation and dysfunction, as IL-6 and IL-8 are known acute-phase signals critical for maintenance and restoration of cellular homeostasis.23 In humans, IL-6 release is a hallmark of inflammation in response to cellular stress. IL-6 activates the Janus kinase (JAK)-STAT signal transduction pathway, thereby promoting host defence.23 Likewise, IL-8 has pleiotrophic functions in endothelial cells, including regulation of neutrophil chemotaxis and crosstalk between coagulation and inflammation mechanisms. IL-8 mediates these effects via activation of adenylyl cyclases, MAPKs and intracellular calcium signalling.24 The results also show that constant moderate hyperoxia increases the release of MIF, an important facilitator of endothelial cell activation mediated by proinflammatory stimuli.25 Hinkelbein et al. investigated the effects of short-term hyperoxia exposure on different organs in rats. They provide evidence that even a relatively short-term hyperoxic exposure can induce significant changes in rat brain protein expression, producing inflammation and injury in rat kidney cells and injury in rat lung cells.26–28 In summary, our in-vitro findings add evidence to the existing hypothesis that hyperoxic O2, either constant or cyclic, causes direct cardiovascular and end-organ inflammation and injury.

Kinases as cellular stress signal transducers

Cellular stress can activate multiple signal transduction pathways that ultimately influence cell survival. Several key survival-associated and cell death–associated signal transduction pathways were phospho-activated by moderate hyperoxia. The MAPKs, particularly p38, were activated under both constant moderate hyperoxia and cyclic hyperoxia/anoxia. p38 is activated by oxidative stress, DNA damage and cytokines, and can lead to apoptosis and necrosis as well as potentiation of the inflammatory response. In contrast, activation of the PI3K/Akt pathway and downstream targets, like mTOR and p70S kinase or PRAS4, may be compensatory mechanisms to support cell survival. These pathways appeared to be selectively activated by cyclic hyperoxia/anoxia rather than constant hyperoxia or anoxia.

The transcription factor STAT6 acts as an independent mediator of antiapoptotic signals by inducing the expression of BCL2L1/BCL-XL.29 In contrast, p53 is a potent inducer of apoptosis, and therefore a counterbalance to STAT6. GSK3 is a downstream element of the PI3Kinase/Akt cell survival pathway. Activity of this pathway can be inhibited by Akt-mediated phosphorylation at GSK3 serines 9 and 21. GSK3 has been linked to apoptosis induced by p53 following DNA damage and oxidative stress.30 Conversely, GSK3 also has antiapoptotic efficacy as it phosphorylates components of the NF-κB system, and inhibition of GSK3 increases TNFα-induced cytotoxicity.31–34 Finally, the metabolic sensor AMPK can also be activated by cellular stress to promote catabolic pathways that generate ATP and inhibit anabolic pathways.35 In many cases, the intracellular signalling cascades are inter-dependent and can either maintain cell growth and survival or induce cell death.

Clinical implications

One rationale for O2 therapy is to compensate for reduced tissue oxygenation in surgical stress, inflammation and sepsis, cardiopulmonary dysfunction, oedema and organ damage. Therefore O2 therapy is often liberally administered to prevent tissue hypoxia. In 2000, a landmark study by Greif et al. demonstrated that the incidence of surgical site infections (SSIs) was reduced from 11.2 to 5.2% in patients who received 80% O2 instead of 30% O2.36 A high O2 partial pressure in surgical wounds increases the production of ROS involved in bactericidal host defences, suggesting that hyperoxia can enhance the ability to fight bacterial infections. Although another large clinical trial by Belda et al. confirmed a net benefit of hyperoxia on fighting infections, other and more recent studies in various patient groups have not supported this hypothesis.37–44

The question of whether hyperoxia is beneficial or deleterious is currently an intensely debated topic in perioperative medicine. A recent Cochrane review by Wetterslev et al.4 found that evidence is lacking for a beneficial effect on SSIs when the fraction of inspired O2 is at least 60%, whereas the risk for adverse events, including mortality, may be increased. Interestingly, a recent post-hoc analysis of the PROXI trial suggests that perioperative hyperoxia may be associated with complications that include MI and acute coronary syndrome.5 Meyhoff et al.2,3 suggested that perioperative hyperoxia may be associated with shorter cancer-free survival in patients undergoing cancer surgery. Likewise, growing evidence from ICUs suggests that arterial hyperoxia might be associated with poor hospital outcome and that more evidence is needed to provide critical care physicians with optimal O2 targets.6,7,45 Recently, the European Resuscitation Council has suggested a more cautious and individually tailored use of O2, especially in newborns and adult patients with myocardial injury.7,8

By demonstrating that even moderate constant hyperoxia and cyclic hyperoxia/anoxia can cause as great or greater endothelial injury than constant anoxia, our study makes a case for the re-evaluation of optimal O2 targets. The results support the concept that perioperative hyperoxia may harm the vascular endothelium. We propose that the tightly regulated micro-environment of the vascular endothelium may be directly disrupted by both hypoxic and hyperoxic exposure, thereby potentially aggravating surgical stress and inflammation and ultimately causing endothelial dysfunction with subsequent cardiovascular and end-organ damage.

Study limitations

Although the HUVEC model is a classic in-vitro technique for the study of human vascular endothelium under controlled conditions, our results need to be replicated in other endothelial cell types.46 It would also be of clinical interest to study the effect of constant moderate hyperoxia or cyclic hyperoxia/anoxia on HUVECs subjected to other models of injury and disease. One is exposure to lipopolysaccharide, as perioperative hyperoxia accompanies surgical stress or bacterial inflammation. We speculate that the injurious effects of constant moderate hyperoxia or cyclic hyperoxia/anoxia would be even more pronounced under these conditions. In addition, selective blockade of activated pathways (e.g. JAK-STAT and MAPK) is required to assess their causal relationship with inflammation and cell death. As release rates of IL-6, IL-8 and MIF were upregulated, it would be interesting to assess the effects of constant moderate and cyclic hyperoxia/anoxia on neutrophil chemotaxis and coagulation.23–25 It is also important to note that ‘constant anoxia’ in our bioreactor may actually be 0 to 1% O2 (severe hypoxia rather than true anoxia).47 This may in part explain the more deleterious effects of hyperoxia. As endothelial cells interact with many other cells, results must be interpreted with caution.16 Therefore, we propose to confirm our findings by determining indices of endothelial injury, for example the endothelial glycocalyx layer and endothelial biomarkers, and inflammation during exposure to altered O2 in the clinical setting.


Constant moderate hyperoxia (40% O2) and cyclic hyperoxia/anoxia had injurious effects on HUVECs, including suppression of cell growth and viability, enhanced apoptosis and necrosis and induction of inflammatory cytokine release. Given the vast surface area of the vascular endothelium in humans, these responses may play important roles in hyperoxia-related morbidity and mortality. Although the phenomenon of endothelial injury after moderate hyperoxic exposure seems of clinical interest, our results should be interpreted with caution and cannot be extrapolated to the clinical setting without further investigation.

Acknowledgements relating to this article

Assistance with the study: none.

Financial support and sponsorship: this study was funded by the Medical Scientific Fund of the Mayor of the City of Vienna, Austria.

Conflicts of interest: none.

Presentation: none.


1. Ray PD, Huang B-W, Tsuji Y. Reactive oxygen species (ROS) homeostasis and redox regulation in cellular signaling. Cell Signal 2012; 24:981–990.
2. Meyhoff CS, Jorgensen LN, Wetterslev J, et al. PROXI Trial Group. Increased long-term mortality after a high perioperative inspiratory oxygen fraction during abdominal surgery: follow-up of a randomized clinical trial. Anesth Analg 2012; 115:849–854.
3. Meyhoff CS, Jorgensen LN, Wetterslev J, et al. PROXI Trial Group. Risk of new or recurrent cancer after a high perioperative inspiratory oxygen fraction during abdominal surgery. Br J Anaesth 2014; 113:i74–i81.
4. Wetterslev J, Meyhoff CS, Jorgensen LN, et al. The effects of high perioperative inspiratory oxygen fraction for adult surgical patients. Cochrane Database Syst Rev 2015; 6:CD008884.
5. Fonnes S, Gögenur I, Sondergaard ES, et al. Perioperative hyperoxia – long-term impact on cardiovascular complications after abdominal surgery, a post hoc analysis of the PROXI trial. Int J Cardiol 2016; 215:238–244.
6. Kilgannon JH, Jones AE, Shapiro NI, et al. Association between arterial hyperoxia following resuscitation from cardiac arrest and in-hospital mortality. JAMA 2010; 303:2165–2171.
7. Kilgannon JH, Jones AE, Parrillo JE, et al. Relationship between supranormal oxygen tension and outcome after resuscitation from cardiac arrest. Circulation 2011; 123:2717–2722.
8. Soar J, Nolan JP, Böttiger BW, et al. Adult advanced life support section Collaborators. European Resuscitation Council Guidelines for Resuscitation 2015: Section 3. Adult advanced life support. Resuscitation 2015; 95:100–147.
9. Nikolaou NI, Arntz HR, Bellou A, et al. Initial management of acute coronary syndromes section Collaborator. European Resuscitation Council Guidelines for Resuscitation 2015 Section 8. Initial management of acute coronary syndromes. Resuscitation 2015; 95:264–277.
10. Baumgardner JE, Markstaller K, Pfeiffer B, et al. Effects of respiratory rate, plateau pressure, and positive end-expiratory pressure on PaO2 oscillations after saline lavage. Am J Respir Crit Care Med 2002; 166:1556–1562.
11. Boehme S, Bentley AH, Hartmann EK, et al. Influence of inspiration to expiration ratio on cyclic recruitment and derecruitment of atelectasis in a saline lavage model of acute respiratory distress syndrome. Crit Care Med 2015; 43:e65–e74.
12. Klein KU, Hartmann EK, Boehme S, et al. PaO2 oscillations caused by cyclic alveolar recruitment can be monitored in pig buccal mucosa microcirculation. Acta Anaesthesiol Scand 2013; 57:320–325.
13. Klein KU, Boehme S, Hartmann EK, et al. Transmission of arterial oxygen partial pressure oscillations to the cerebral microcirculation in a porcine model of acute lung injury caused by cyclic recruitment and derecruitment. Br J Anaesth 2013; 110:266–273.
14. Wu J, Hafner C, Schramel JP, et al. Cyclic and constant hyperoxia cause inflammation, apoptosis and cell death in human umbilical vein endothelial cells. Acta Anaesthesiol Scand 2015; 60:492–501.
15. Wojta J, Zoellner H, Gallicchio M, et al. Gamma-interferon counteracts interleukin-1 alpha stimulated expression of urokinase-type plasminogen activator in human endothelial cells in vitro. Biochem Biophys Res Commun 1992; 188:463–469.
16. Lehle K, Straub RH, Morawietz H, et al. Relevance of disease- and organ-specific endothelial cells for in vitro research. Cell Biol Int 2010; 34:1231–1238.
17. Wu J, Stefaniak J, Hafner C, et al. Intermittent hypoxia causes inflammation and injury to human adult cardiac myocytes. Anesth Analg 2015; 122:373–380.
18. Vane JR, Anggård EE, Botting RM. Regulatory functions of the vascular endothelium. N Engl J Med 1990; 323:27–36.
19. Lee WL, Slutsky AS. Sepsis and endothelial permeability. N Engl J Med 2010; 363:689–691.
20. Cahill PA, Redmond EM. Vascular endothelium – gatekeeper of vessel health. Atherosclerosis 2016; 248:97–109.
21. Jaffe EA. Cell biology of endothelial cells. Hum Pathol 1987; 18:234–239.
22. Michiels C, Arnould T, Remacle J. Endothelial cell responses to hypoxia: initiation of a cascade of cellular interactions. Biochim Biophys Acta 2000; 1497:1–10.
23. Hunter CA, Jones SA. IL-6 as a keystone cytokine in health and disease. Nat Immunol 2015; 16:448–457.
24. Bhandari V, Elias JA. Cytokines in tolerance to hyperoxia-induced injury in the developing and adult lung. Free Radic Biol Med 2006; 41:4–18.
25. Tillmann S, Bernhagen J, Noels H. Arrest functions of the MIF ligand/receptor axes in atherogenesis. Front Immunol 2013; 4:115.
26. Hinkelbein J, Feldmann RE Jr, Kalenka A. Time-dependent alterations of cerebral proteins following short-term normobaric hyperoxia. Mol Cell Biochem 2010; 339:9–21.
27. Hinkelbein J, Bohm L, Spelten O, et al. Hyperoxia-induced protein alterations in renal rat tissue: a quantitative proteomic approach to identify hyperoxia-induced effects in cellular signaling pathways. Dis Markers 2015; 2015:964263.
28. Spelten O, Wetsch WA, Wrettos G, et al. Response of rat lung tissue to short-term hyperoxia: a proteomic approach. Mol Cell Biochem 2013; 383:231–242.
29. Hebenstreit D, Wirnsberger G, Horejs-Hoeck J, et al. Signaling mechanisms, interaction partners, and target genes of STAT6. Cytokine Growth Factor Rev 2006; 17:173–188.
30. Watcharasit P, Bijur GN, Zmijewski JW, et al. Direct, activating interaction between glycogen synthase kinase-3beta and p53 after DNA damage. Proc Natl Acad Sci U S A 2002; 99:7951–7955.
31. Hoeflich KP, Luo J, Rubie EA, et al. Requirement for glycogen synthase kinase-3beta in cell survival and NF-kappaB activation. Nature 2000; 406:86–90.
32. Schwabe RF, Brenner DA. Role of glycogen synthase kinase-3 in TNF-alpha-induced NF-kappaB activation and apoptosis in hepatocytes. Am J Physiol Gastrointest Liver Physiol 2002; 283:G204–G211.
33. Demarchi F, Bertoli C, Sandy P, et al. Glycogen synthase kinase-3 beta regulates NF-kappa B1/p105 stability. J Biol Chem 2003; 278:39583–39590.
34. Schotte P, Van Loo G, Carpentier I, et al. Lithium sensitizes tumor cells in an NF-kappa B-independent way to caspase activation and apoptosis induced by tumor necrosis factor (TNF). Evidence for a role of the TNF receptor-associated death domain protein. J Biol Chem 2001; 276:25939–25945.
35. Hardie DG, Schaffer BE, Brunet A. AMPK: an energy-sensing pathway with multiple inputs and outputs. Trends Cell Biol 2016; 26:190–201.
36. Greif R, Akça O, Horn EP, et al. Outcomes Research Group. Supplemental perioperative oxygen to reduce the incidence of surgical-wound infection. N Engl J Med 2000; 342:161–167.
37. Belda FJ, Aguilera L, García de la Asunción J, et al. Supplemental perioperative oxygen and the risk of surgical wound infection: a randomized controlled trial. JAMA 2005; 294:2035–2042.
38. Pryor KO, Fahey TJ, Lien CA, et al. Surgical site infection and the routine use of perioperative hyperoxia in a general surgical population: a randomized controlled trial. JAMA 2004; 291:79–87.
39. Meyhoff CS, Wetterslev J, Jorgensen LN, et al. Effect of high perioperative oxygen fraction on surgical site infection and pulmonary complications after abdominal surgery: the PROXI randomized clinical trial. JAMA 2009; 302:1543–1550.
40. Kao LS, Millas SG, Pedroza C, et al. Should perioperative supplemental oxygen be routinely recommended for surgery patients? A Bayesian meta-analysis. Ann Surg 2012; 256:894–901.
41. Togioka B, Galvagno S, Sumida S, et al. The role of perioperative high inspired oxygen therapy in reducing surgical site infection: a meta-analysis. Anesth Analg 2012; 114:334–342.
42. Duggal N, Poddatoori V, Poddatorri V, et al. Perioperative oxygen supplementation and surgical site infection after cesarean delivery: a randomized trial. Obstet Gynecol 2013; 122:79–84.
43. Williams NL, Glover MM, Crisp C, et al. Randomized controlled trial of the effect of 30% versus 80% fraction of inspired oxygen on cesarean delivery surgical site infection. Am J Perinatol 2013; 30:781–786.
44. Kurz A, Fleischmann E, Sessler DI, et al. Effects of supplemental oxygen and dexamethasone on surgical site infection: a factorial randomized trial. Br J Anaesth 2015; 115:434–443.
45. Helmerhorst HJF, Roos-Blom M-J, van Westerloo DJ, et al. Association between arterial hyperoxia and outcome in subsets of critical illness: a systematic review, meta-analysis, and meta-regression of cohort studies. Crit Care Med 2015; 43:1508–1519.
46. Onat D, Brillon D, Colombo PC, et al. Human vascular endothelial cells: a model system for studying vascular inflammation in diabetes and atherosclerosis. Curr Diab Rep 2011; 11:193–202.
47. Ward JPT. Oxygen sensors in context. Biochim Biophys Acta 2008; 1777:1–14.
© 2017 European Society of Anaesthesiology