Ischemia–reperfusion (IR) features massive oxidative stress of tissues and cytokine response. Propofol and sevoflurane, both of which are commonly used anesthetics, are thought to have different antioxidant activities. The aim of this study is to delineate the influence of these two drugs on the production of free radicals and proinflammatory cytokines in IR conditions via in vitroand in vivo models.
An in vitro IR model was performed by incubating porcine cells (including mononuclear cells, and coronary and aortic smooth muscle cells) with either propofol 25 μM or sevoflurane 2% in the hypoxia chamber (1% O2, 37°C) for 1 hour, followed by room temperature air for 2 hours. Reactive oxygen species (ROS) and tumor necrosis factor-α (TNF-α) were also measured via flow cytometry and enzyme-linked immunosorbent assay methods, resp
The results showed significant reduction of both ROS and TNF-α levels in the propofol group in vitro IR model. However, there was no difference in lipid peroxidation and TNF-α level between propofol and sevoflurane for the in vivoIR model.
We concluded that propofol, compared with sevoflurane, can significantly inhibit ROS formation on a cell level. In addition, propofol can significantly inhibit TNF-α formation of monocytes and coronary smooth muscle cells but not aortic smooth muscle cells.
free radicalsischemia–reperfusion injurypropofolsevofluranetumor necrosis factor-α;
Ischemia–reperfusion (IR) produces a lot of free radicals and causes tissue damage.1 In clinical practice, there are lots of conditions associated with IR, such as tourniquet-induced IR in peripheral or major vascular surgery or organ transplantation. When IR occurs, overproductive reactive oxygen species (ROS) and inflammatory cytokines such as tumor necrosis factor-α (TNF-α) may cause severe problems owing to associated tissue damage. It has been said that vascular smooth muscle cells are the possible target.2Meanwhile, modern anesthesia uses rapid recovery anesthetic drugs to reduce the postanesthesia effect. Propofol and sevoflurane are two of the commonly used drugs. However, it has been reported that propofol has a more antioxidant effect and reduces free radical-induced injury.3, 4 Propofol was also reported for its proinflammatory cytokine response suppression in sepsis.4, 5 However, there are little data available that indicate the difference of these two drugs in an IR model. This study investigated the antioxidant and anti-inflammatory effects of propofol and sevoflurane via in vitro and in vivo IR models with porcine cells and animal methods. As we know, porcine cells are even closer to human cells than rodents. The results should provide better clinical implications.
2.1. Porcine polymorphonuclear cells and smooth muscle cells preparation
Peripheral porcine blood polymorphonuclear (PMN) cells were purified with Ficoll-Paque density gradient centrifugation. Whole blood (50 mL) was diluted with an equal volume of Hank's Balanced Salt Solution (HBSS) with 5% bovine serum albumin and then carefully layered over Ficoll-Paque. After centrifugation was performed (40 minutes, 400g), the PMN layer was carefully aspirated, washed in HBSS three times, and incubated in 60 mm plastic culture dishes in HBSS with 5% bovine serum albumin. The nonadherent cells were aspirated and the adherent cells (PMN) were incubation for 3 days for stability before experimentation. Porcine aorta smooth muscle cells (PAOSMC, Cell Application Inc., San Diego, CA, USA) and coronary smooth muscle cells (PCASMC, Cell Application Inc.) were all cultured in 37°C 5% CO2/95% air, and maintained with specific growth medium (P311-500) without fetal bovine serum added. Cells were stored in liquid nitrogen if not used. Medium was changed every other day until the culture was > 60% confluent and the subculture was 80% confluent.
2.2. Experimental protocol in cells and hypoxic cytotoxicity assay
Cells were coincubated with either propofol 25 μM or sevoflurane 2% condition and then incubated in a hypoxia culture chamber (1% O2, 37°C) for 1 hour and reoxygenated (21% O2, 37°C) for 2 hours. Cells were then collected and ROS and proinflammatory cytokines were measured. Cell damage evaluation after hypoxia was evaluated by 2-(4-iodo-phenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium (WST-1) colorimetric method.6 WST-1, a water-soluble formazan enzyme, is thought to reflect a sensitive respiratory enzyme activity after injury. Cells (5 × 103/well) were in microplates (96 wells, flat bottom) with 100 μL medium and 10 μL WST-1 reagent (Roche Molecular Biochemicals, Indianapolis, IN, USA) was added and then incubated for 4 hours. Optical density was measured at 450 nm on a microplate reader.
2.3. Superoxide measurement in cells
Cellular accumulation of superoxide was determined with the dihydroethidium method. To measure ROS levels, treated cells were incubated with 1M of dihydroethidium for 30 minutes at 37°C, followed by cell washing and resuspension in ice-cold phosphate-buffered saline. Cell Quest software was applied for flow cytometric analysis (FACS Calibur, BD Biosciences, San Jose, CA, USA). The mean fluorescence intensity of white cell, separated with electronic gating in the forward scatter/side scatter dot plot, was measured and quantified in arbitrary units of 10,000 events.
2.4. Animal preparation
The Institutional Animal Ethic Committee in our university (College of Medicine, National Cheng Kung University, Tainan, Taiwan) approved all procedures in this study. Pigs (55–65 kg, Duroc), purchased from Livestock Research Institute (Council of Agriculture, Hsinhua, Taiwan), were randomly assigned to propofol or sevoflurane groups (n = 5 in each groups). Xylazine (0.2 mg/kg, intramuscular) was used for sedation. Thiopental (25 mg/kg) and succinylcholine (1 mg/kg) were administrated intravenously via the ear vein for anesthetic induction and tracheal intubation. Carotid and femoral artery catheters were then inserted for blood sampling and represent upper and lower body's conditions, respectively. Arterial blood pressure, large bore intravenous catheter, electrocardiography, end-tidal CO2, and pulmonary artery catheter were also used for intraoperative monitor.
2.5. Animal experimental protocol
Animals were maintained with propofol (10–15 mg/kg/h) or sevoflurane (1.5–2.5%) according blood pressure change. Serial blood serum sampling (10 mL for each time point) including basal level, 30 minutes, and 60 minutes after aortic clamp and 60 minutes and 120 minutes after aortic declamp were also harvested when surgical procedure for special designed ventricular assist device implantation was performed in the descending thoracic aorta. During aorta clamping or declamping, vessel control drug (nitroglycerin and epinephrine) plus fluid infusion were administered to maintain blood pressure within ±25% of the basal level. The total clamp time (clamp over descending thoracic aorta) was 1 hour and blood pressure was also recorded. Serum, centrifuged from whole blood (14,000 rpm, 4°C, 15 minutes), was then stored at −80°C for analysis.
2.6. Lipid oxidation measurement
The antioxidant effect of propofol on lipid peroxidation was evaluated by determining malondialdehyde (MDA) level in the blood via the thiobarbituric acid reacting substances method.7 Serum mixed with trichloracetic acid and thiobarbituric acid was boiled. Butanol was then added to the tube and results were obtained at 532 nm after centrifugation.
2.7. Proinflammatory cytokine analysis
The level of proinflammatory cytokines (TNF-α) in serum and culture medium were analyzed with a commercial enzyme-linked immunosorbent assay kit (R&D Systems, UK) and obtained with an enzyme-linked immunosorbent assay reader at 570 nm. Briefly, after centrifugation, supernatant were collected and stored at −80°C. Cytokine levels were determined by interpolation with standard curves assayed on individual plates and normalized to protein content in each sample.
2.8. Statistical analysis
The results are expressed as means ± standard error mean. One-way analysis of variance and Student t test or multiple comparisons were used for the statistical evaluation of differences among means. A value of p < 0.05 was considered to be statistically significant.
The hypoxia injury evaluations for the in vitro IR model with different anesthetics within different porcine cells were shown in Figure 1. Cellular IR model damaged all cell types regardless of the presence of anesthetics [all cell data were presented in a serial of monocytes, PAOSMC, and PCASMC; control (C), 1.08 ± 0.03, 1.07 ± 0.02, and 1.00 ± 0.11; hypoxia (H), 0.62 ± 0.05, 0.37 ± 0.06, and 0.70 ± 0.06; hypoxia plus propofol (H + P), 0.62 ± 0.04, 0.35 ± 0.06, and 0.62 ± 0.09; and hypoxia plus sevoflurane (H + S), 0.59 ± 0.05, 0.45 ± 0.04, and 0.66 ± 0.08; n = 4 in each group, * showed t < 0.05 compared with C]. Sevoflurane showed a little increase than propofol but the difference is not statistically insignificant. The production ratio of ROS in different hypoxia conditions is shown in Figure 2 (C, 0.56 C,0.03, 0.54 .00.04, and 0.57 .50.06; H, 1.03, 0.02, 1.00 .00.10, and 1.00 .00.04; H + P, 0.65, 0.02, 0.62 .00.02, and 0.60 .60.08; and H + S, 1.10 ± 0.07, 0.88 ± 0.02, and 0.88 ± 0.06; n = 4 in each group, and *, **, shows p < 0.05 compared with C, H + P group, respectively, as shown in Figures 3A and 3B). Propofol significantly reduced the production of ROS compared with H + S group. In Figure 3, the level of TNF-α was also significantly inhibited by propofol in monocytes (Figure 3A; monocyte, C: 43.56 ± 07.50; H: 1128.3 ± 276.4; H + P: 53.9 ± 7.2; and H + S: 1553.4 ± 107.2; n = 4 in each group) and coronary smooth muscle cells but not in aorta smooth muscle cells (Figure 3B; PAOSMC, C: 34.5 ± 20.7; H: 61.0 ± 19.5; H + P: 65.8 ± 22.4; and H + S: 89.0 ± 23.7; and PCASMC, C: 32.8 ± 6.8; H: 74.6 ± 16.4; H + P: 28.7 ± 6.9; and H + S: 42.0 ± 12.6; n = 4 in each group). Arterial smooth muscle cells showed a significant difference of TNF-α production than monocytes but similar on ROS production in Figure 2. Furthermore, different arterial smooth muscle cell also showed different TNF-α production in an IR condition in the presence of different anesthetics but the difference is subtle compared with monocytes. It also hints that monocytes are the possible target of propofol and different cells present different effects under the same concentration of anesthetics.
Download full-size image
Figure 1. This figure shows an in vitro ischemia–reperfusion injury with the 2-(4-iodo-phenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)- 2H-tetrazolium method. It shows that the in vitro ischemia–reperfusion model induced cell damage whether anesthetics were used or not. There was no difference in various cells or even a slight increase in the sevoflurane group without statistical significance (n = 4 in each group, * shows t < 0.05 compared with control). C = control; H = hypoxia group; H + P = hypoxia plus propofol group; H + S = hypoxia plus sevoflurane group; PAOMSC = porcine aorta smooth muscle cells; PCAMSC = porcine coronary smooth muscle cells.
Download full-size image
Figure 2. This figure shows the ratio of reactive oxygen species in different cells treated with sevoflurane and propofol under in vitro ischemia–reperfusion conditions. Propofol significantly suppressed the reactive oxygen species formation compared with other groups (n = 4 in each group, *, **, shows p < 0.05 compared with control, hypoxia group, respectively). C = control; H = hypoxia group; H + P = hypoxia plus propofol group; H + S = hypoxia plus sevoflurane group; PAOMSC = porcine aorta smooth muscle cells; PCAMSC = porcine coronary smooth muscle cells.
Download full-size image
Figure 3. Tumor necrosis factor-α production in different cells under in vitro ischemia–reperfusion conditions. Propofol significant suppressed tumor necrosis factor-α production in: (A) monocytes and in coronary smooth muscle cells; but not in (B) aortic smooth cells (n = 4 in each group, *, ** shows p < 0.05 compared with control, hypoxia groups, respectively). C = control; H = hypoxia group; H + P = hypoxia plus propofol group; H + S = hypoxia plus sevoflurane group; PAOMSC = porcine aorta smooth muscle cells; PCAMSC = porcine coronary smooth muscle cells.
In an animal IR model, MDA production was significantly suppressed in the H + P groups compared with the H + S groups after 60-minutes thoracic aorta clamping (Figure 4; B: basal level after anesthesia induction. C30, C60: 30 minutes or 60 minutes later after aortic clamping, D 60, D120: 60 minutes or 120 minutes later after aortic declamping; Prop-U, Prop-L, Sevo-U, Sevo-L: proximal or distal aortic clamp sampling with propofol or sevoflurane maintenance; MDA level, Sevo-U: B 28.4 DA0.24, C60 42.8 C62.2, p <0.05 and Sevo-L: B 28.6 <00.28, C60 43.2. 1.5, p <0.05, Prop-U vs. Sevo-U in C60: 34.4 <00.62 vs.42.8 .62.2, p <0.05, Prop-L vs. Sevo-L in C60: 32.9 <01.50 vs.43.2 vs1.5, p <0.05; n = 5 in each group). However, we only demonstrated the elevation of TNF-αtafter aortic clamping for 60 minutes but no difference between different anesthetics in TNF-α production (Figure 5; B vs. C60 for Prop-U, Prop-L, Sevo-U, Sevo-L: 52.9 v3.6 vs. 84.7 vs5.5, 51.1 513.1 vs. 83.8 vs4.0, 59.9 900.8 vs. 85.8 .84.7, and 58.8 .81.6 vs. 81.3 vs6.1, all p <0.05, n = 5 in each group). For comparing the difference between the proximal and distal portion of aortic clamp, either the level of MDA or the level of TNF-α showed no statistic difference in every sampling point.
Download full-size image
Figure 4. This figure expresses the change of malondialdehyde in both groups and upper/lower body. B represents the basal level after anesthesia. C30 and C60 represents 30 minutes or 60 minutes after aortic clamp, respectively. D60 and D120 represented 60 minutes or 120 minutes after aortic declamping respectively (n = 5 in both groups, *, **, shows p < 0.05 compared with basal level and sevoflurane group, respectively).
Download full-size image
Figure 5. The figure expressed the change of tumor necrosis factor-α in both groups and upper/lower body. B represents the basal level after anesthesia. C30 and C60 represents 30 minutes or 60 minutes after aortic clamp, respectively. D60 and D120 represents 60 minutes or 120 minutes after aortic declamp, respectively (n = 5 in both groups, and * shows p < 0.05 compared with basal level).
Both propofol and sevoflurane are commonly used anesthetics due to their fast recovery profile, and they are compared in many different clinical and basic researches. Some researches demonstrated that sevoflurane rather than propofol preserved myocardial function in coronary surgery.8, 9 It has also been said that sevoflurane attenuates the hemodynamic response in reperfusion injury more than propofol.10 However, in other studies, propofol was associated with less of an increase in the level of MDA on tourniquet knee surgery11 and provided lower blood concentration of myeloperoxidase and TNF-α (tissue damage markers) in a piglet IR model.12 It was also indicated that sevoflurane cannot reduce the level of postoperative troponin I in cardiac surgery.13 Compared with propofol, sevoflurane caused no difference in intestine tissue examination after IR injury.10 However, propofol has been reported to reduce circulating catecholamines more than sevoflurane in brain ischemia rats14 and reduce interstitial glycolysis metabolites in tourniquet-induced IR surgery.15 Moreover, Jakobsen et al16 demonstrated that propofol had a benefit for severe preoperative ischemic stress and less incidence of postoperative atrial fibrillation in their 10,535 cardiac surgical procedures.
Sevoflurane and propofol both provided some beneficial effects on the heart. However, there are no data of these two drugs on vascular smooth muscle cells or monocytes for in vitro IR models. For monocytes, propofol reduced interleukin-2 production while stimulated with phyetohaemagglutinin.17 In our study, propofol had better effects on ROS suppression on monocyte and smooth muscle cells via an in vitro IR injury model. Also, propofol had a significant suppression on the production of TNF-α in monocyte and coronary smooth cells but not in aortic smooth muscle cells. Runzer et al18revealed the difference of antioxidant capacity of propofol in different organs. But little data indicated the difference in different cells. There were also no data focusing on the mechanism of those differences. Further detailed studies may be needed to solve this issue.
ROS induced by pathological conditions could produce vascular injury via interacting with monocytes or vascular smooth muscle cells.19 ROS-associated oxidative damage was also thought to be a mediator of vascular injury in hypertension, hyperlipidemia, or diabetes.20 ROS and TNF-α were also involved in the monocyte-induced vascular inflammation in vascular smooth muscle cells.21 It is important to evaluate the influence of different anesthetics in an IR condition on the production of ROS and TNF-α. However, it also needs more surveys to demonstrate the beneficial effects of anesthetics on vascular inflammation.
We failed to demonstrate the difference of MDA or TNF-α production in a whole porcine IR model. We thought that the variable ischemia time and the sampling point may cause this inconsistent result. Annecke et al22 showed that there was no difference of those two drugs in the severity of lung injury with a 90-minute porcine IR model. Papakostas et al23 found that 45-minute thoracoabdominal aortic occlusion can induce mild pancreatic injury. Juel et al24 also presented the intestinal injury in their 60-minute IR porcine model. Except the different IR injury induced by different ischemic times, none of the results in the above studies demonstrated the serum level of ROS or TNF-α in their IR model. Another possible explanation for our negative result is the difference may be demonstrated only in a regional level but not in a whole body serum level. The negative result for the in vivo model may be also for the inappropriate sampling timing. The surge of ROS and proinflammatory cytokines might be highest just after the aortic declamping moment, but we sampled the serum 1 hour later after aortic declamping. The critical sampling point might be missed. This point of view is supported by Aldemr et al's25 study which indicated that propofol reduced ROS injury markers in blood only in the initial 5-minute period in their tourniquet-induced IR model.
In conclusion, on a cell level, compared with sevoflurane, propofol significantly inhibited ROS formation. In addition, propofol also significantly inhibited TNF-α formation of monocyte and coronary smooth muscle cells but not of aortic smooth muscle cells. However, in a whole animal model, the beneficial effect of propofol was not observed 1 hour later after aorta declamping.
The authors would like to thank Alice Y.W. Chang and Julie Y.H. Chan for their helpful and important commentary and assistance for this work. The authors also want to give thanks to Chein-Chi Huang and Pao-Yen Lin for their help in the cell culture and animal study, respectively. This work was also partially supported by research grants from the National Cheng Kung University Hospital (NCKUH95-86), Tainan, Taiwan.