Deleterious effects of oxygen during extracorporeal circulation for the microcirculation in vivo q

Objective : Clinical complications arising from extracorporeal circulation (ECC) have been linked to disturbances in the microcirculation. Hyperoxia, a mainstay of supportive treatment, is clinically used for a variety of pathological states. In previous in vivo animal experiments we found increased leukocyte/endothelial (L/E) cell interaction following ECC due to oxygen derived free radicals. This study was carried out to investigate the link between arterial pO 2 during ECC and the potential damage to the microcirculation, supposedly caused by oxygen derived radicals. Methods : Intravital ﬂuorescence microscopy was used on the dorsal skinfold chamber preparation in syrian golden hamsters. ECC was introduced via a micro-rollerpump (0.7 ml/min) and a 60 cm silicon tube (1 mm inner diameter) shunted between the carotid artery and the jugular vein after application of 300 IE Heparin/kg/bw. Experiments were performed in chronically instrumented, awake animals (age: 10–14 weeks, weight: 65–75 g). Control inspired room air, experimental group 1 inspired 100% oxygen, group 2 received 100% oxygen and 2000 IE of Heparin i.v. ( n ¼ 7/group), that releases endothelial bound superoxide dismutase, a natural scavenger of oxygen derived free radicals in the hamster. Results : Normobaric inhalation of 100% oxygen increased arterial pO 2 from 64 ^ 8.1 mmHg to 512 ^ 124 mmHg ( P , 0 : 05 vs. baseline). ECC under 100% oxygen reduced functional capillary density (FCD) to 70% of baseline values 8 h after ECC ð P , 0 : 05 Þ : Adherent leukocytes in postcapillary venules and arterioles increased signiﬁcantly ð P , 0 : 05 Þ : 2000 IE Heparin prevented the reduction in FCD and decreased the number of adherent leukocytes. Conclusions : Reduction in FCD, increased leukocyte adherence to the microvascular endothelium of postcapillary venules and arterioles under hyperoxia compared to ECC under room air conditions, demonstrates harmful effects of oxygen during ECC in vivo. A high dose of Heparin enhances functional capillary density, thus attenuating the microvascular dysfunction/damage in the period after ECC.


Introduction
Although there is some acceptance that oxygen supplemention is beneficial to augment tissue oxygenation when used in states of anemia or ischemia due to occlusive vessel disease [3,10], no evidence exists that a positive clinical outcome is warranted. One possible reason is vasoconstriction, caused by hyperoxia, thus lowering tissue perfusion as demonstrated extensively in vitro (isolated vessels) and in vivo [5,6]. Recently, using intravitalmicroscopy, Tsai et al. found a decrease of functional capillary density during hyperoxia [24]. In the context of cardiopulmonary bypass (CPB) less information is available about potential effects of hyperoxemia to improve tissue oxygenation, however, adverse events have been reported [14].
We hypothesized that hyperoxia could aggravate extracorporeal circulation (ECC) induced injury to the microcirculation due to increased generation of reactive oxygen species (ROS), which may be partially responsible for the systemic activation of cellular and humoral components during cardiopulmonary bypass (CPB), extracorporeal membrane oxygenation (ECMO) or hemodialysis [9]. This inflammatory response is represented by clinical complications from moderate disturbances of organ functions ranging from increased body temperature European Journal of Cardio-thoracic Surgery 26 (2004) 564-570 www.elsevier.com/locate/ejcts or leukocytosis to severe dysfunction of single organs up to multi-organ failure [11]. Neutrophils play a pivotal role because of their ability to release oxygen derived free radicals and proteolytic enzymes with subsequent tissue damage [12], followed by upregulation of leukocyte adhesion molecules (CD11/CD18) and expression of adhesion molecules on endothelium. In previous in vivo experiments we demonstrated that exposition of blood to an extracorporeal circuit induces dysregulation in the microcirculatory compartment [15]. Up to date no information is available concerning arterial pO 2 during ECC in regard to a probable damage of the microcirculation. The present study was carried out to determine the extent by which hyperoxic blood may interfere with ECC induced microcirculatory disturbances in the unanaesthetized hamster skinfold model and if formation of oxygen free radicals is involved.

Model
Experiments were carried out in male Syrian golden hamsters (age: 10 -14 weeks, weight: 65 -75 g). The surgical technique has previously been described in detail [8]. Briefly, the dorsal skin fold, consisting of two layers of skin and muscle tissue, was fitted with two titanium frames with a 15-mm circular opening surgically installed under Narcoren (60 mg/kg body wt., Pentobarbital, Merial, Hallbergmoos, Germany) anaesthesia. Layers of skin muscle were carefully separated from subcutaneous tissue and removed until a thin monolayer of muscle and one layer of intact skin remained. A cover glass held by one frame was placed on the exposed tissue, allowing intravital observation of the microvasculature. The second frame remained open, exposing intact skin.
Catheters were implanted in the jugular vein (polythene tubing 0.40 £ 0.80 mm) and the carotid artery (polythene tubing 0.28 £ 0.61 mm). All experiments were performed after a recovery period of at least 24 h after catheter implantation and 3.6^0.6 days (mean^SD) after chamber implantation. Chambers, catheters and instruments were sterilized before use, surgery was performed under aseptic conditions. Five preparations with signs of inflammation, edema, bleeding spots and no flow were excluded from further investigations. Inclusion criteria for systemic parameters were: systemic mean arterial blood pressure (MABP) greater than 90 mmHg and hematocrit over 45%. All animals received humane care in compliance with the European Convention on Animal Care and the study was approved by the institutional and regional committee for animal care.

Intravital microscopy
Microscopic observations were performed using an intravital microscope (Leica DMLM, Wetzlar, Germany) with a 20 £ SW 0.40 BD NA objective (Leica Fluotar, Wetzlar, Germany). A 100-W Hg light source was used for epi-illumination. Contrast enhancement for transillumination was accomplished with a blue filter (420 nm), which selectively passes light in the maximum absorption band of hemoglobin, causing red blood cells to appear as dark objects in an otherwise gray background. A heat filter was placed in the light path prior to the condenser. Microscopic images were viewed by a closed circuit video system consisting of a CCD camera (Kappa CF 8/4 NIR, Gleichen, Germany) and a monitor (Sony, Japan).
Leukocytes were stained with rhodamine 6 G (Sigma, St. Louis, MO) and classified by fluorescence microscopy according to their interaction with the endothelial lining as adherent, rolling or free flowing cells. Adherent leukocytes were defined in each vessel segment as cells that did not move or detach from the endothelial lining within an observation period of 30 s and are expressed as number of cells per square millimeter (cells/mm 2 ) of vessel surface as calculated from diameter and length (100 mm) of the vessel segment studied. Rolling leukocytes are expressed as percentage of non-adherent leukocytes passing through the observed vessel segment within 30 s.
Red blood cell (RBC) velocity and diameters were measured in 5 venules per observation chamber using a computer-assisted analysis system (Capimage, Dr Zeintl Ingenieurbuero, Heidelberg, Germany).
Functional capillary density (FCD) was assessed in nine successive microscopic fields by transillumination in a region of approximately 1.3 mm 2 ; the initial field was chosen where microvessels were in focus and there were between two and five capillaries with RBC flow in the field of view. Systematic observations were achieved by displacing the microscopic field of view in three consecutive steps in the lateral direction (relative to the observer). Each step had the extension of the area seen on the monitor, being 436 mm long when referred to the tissue. The same procedure was repeated after moving the image by one microscopic field in the vertical (333 mm) direction. FCD was evaluated by measuring the length of capillaries that had red blood cell flow. A capillary was defined to be functional if passing RBCs were noted within a 20-s observation period.
All intravital microscopic observations (on-line) were recorded and evaluated later (off-line) in order to minimize the duration of the experiment and the time of light exposure to the tissue.
Arterial blood pressure and heart rate were monitored via carotid artery connected to a transducer (Bently, Uden, Holland) with an analog recording system (Hellige, Freiburg, Germany) for continuous measurements.

Extracorporeal circulation (ECC)
ECC was introduced via a micro-rollerpump (ISM 833A, Ismatec, Glattbrugg-Zürich, Switzerland) and a silastic tube (1 mm inner diameter, length 60 cm, Migge, Heidelberg, Germany) shunted between the carotid artery and the jugular vein. The sterilized extracorporeal circuit was primed with 1 ml solution and a flow rate of 0.7 ml/min was used. Animals tolerate these procedures and show no signs of discomfort or significant changes in blood pressure or heart rate. The expected cardiac output of hamsters is 461^29 ml/min/kg/body weight [7]. The percentage of the cardiac output represented by the ECC is 3% per minute, but within 30 min 400% of the total hamster blood volume (5.25 ml/70 g/body weight) contacted the ECC-system.

Oxygen inspiration
Inspiratory gas was made by mixing room air and continuous pure oxygen gas. A closed tube system and a jet ventilator (Venturi-principle) warranted standardized conditions. Oxygen concentration was monitored in situ by means of an oxygen monitor (Greisinger Electronic GmbH, Regenstaufen, Germany).

Blood gas analysis
Blood gas analysis was performed in all animals at BL, 30 min and 4 h using 120 ml full blood (ABL 500, Radiometer Copenhagen, Denmark).

Experimental protocol
Twenty one animals undergoing 30 min of isovolemic ECC were studied in a randomised way. After intravenous application of rhodamine 6 G (2 mg/40 ml) and heparin (300 IU/kg body wt. in 200 ml Ringer's solution; Braun AG, Melsungen, Germany) baseline measurements of FCD, vessel diameter, RBC velocity and leukocyte/endothelium interaction were performed. MABP and heart rate were monitored and blood samples were collected from the carotid artery catheter. Follow up measurements were performed at time points (TP) 30 min, 4 and 8 h. Control group ðn ¼ 7Þ received room air. In experimental group I ðn ¼ 7Þ 100% oxygen was applied, in Group II ðn ¼ 7Þ 100% oxygen and 2000 IU Heparin/kg/body weight was used.

Statistical analysis
Data were analysed using SPSS w -statistical analysing software (SPSS Software Inc. Chicago, USA) using ANOVA, t-test or Wilcoxon test with Bonferroni correction. Changes were deemed statistically significant for P , 0:05:

Leukocyte/endothelial cell interaction in arterioles
Under baseline conditions the majority of leukocytes did not interact with endothelial cells in arterioles. After 100% oxygen inhalation, sticker increased at tp 8 h (*P ¼ 0:026 tp vs. baseline; #P ¼ 0:017 control vs. group I). In group II the increase of sticker was less pronounced (Fig. 3). It was not possible to detect roller in arterioles.

Microhemodynamic parameters
Diameter of postcapillary venules and blood cell velocity in postcapillary venules were neither between time points nor between experimental groups and control  (Table 1). At tp 4 h RBC velocity in arterioles was higher, when compared to control ðP , 0:05Þ:

Macrohemodynamic parameters
Arterial bloodpressure and heartrate were stable and not significantly affected by ECC or treatment. During the time of the experiment no states of low pressure were observed.

Blood analysis
Analysis of partial oxygen pressures (pO 2 ) in arterial blood demonstrated stable conditions throughout the experiment in all groups. Under baseline conditions during room air breathing (control) pO 2 was 64^8.1 mmHg compared to 512^124 mmHg (Group 1 means^SD, P , 0:05 vs. baseline) and 502^116 mmHg (Group 2; means^SD, P , 0:05 vs. baseline).

Anticoagulation
Application of 300 as well as 2000 IU/kg/body wt. heparin resulted in a prolongation of partial thromboplastin time (PTT) to more than 300 s at the end of ECC compared to a baseline level of 26.5^5.9 s ðn ¼ 3Þ:

Discussion
The principal findings of this study are that (1) normobaric inspiration of 100% oxygen significantly reduced functional capillary density (FCD) and (2) increased rolling and sticking in postcapillary venules and arterioles. (3) Treatment with high dose heparin reduced L/E cell interaction and prevented severe reduction of FCD.
Administration of high concentrations of oxygen (hyperoxia) is a mainstay of clinical supportive treatment for a variety of pathologies. However, hyperoxia, by generating systemic reactive oxygen species (ROS), can exacerbate organ failure by causing cellular injury and damage solid organs like the kidneys or the brain [21,25]. Effects of hyperoxia on the lung include pulmonary edema, inflammation, and respiratory failure [2]. In a clinical study with 19 patients undergoing CPB for coronary heart disease hyperoxic reoxygenation was associated with increased cardiac Malondialdehyde (MDA) levels immediately after aortic declamping, emphasizing the importance of the adequate oxygenation level [1]. During surgical correction of congenital cardiac defects, the use of hyperoxic CPB might be harmful for the hypoxic immature heart [13].
In previous studies we demonstrated in vivo that ECC induces L/E cell interaction and that the intensity of induction was ECC time dependent and related to the production of oxygen derived free radicals [15]. The consequences for the microcirculation are demonstrated in our study. The finding that FCD is reduced after normobaric hyperoxia is in accordance with a report from others, who found a reduced capillary flow in the pig brain [23]. Tsai et al. explained this phenomenon in hamsters with vasoconstriction of arterioles. Notably they did not detect a reduction in oxygen delivery in the microcirculation using the Pd-phosphorescence quenching method [24], suggesting a compensatory mechanism, which is regulated upstream of the capillary level. Big Endothelin 1 (Big-ET-1), a potent vasoconstrictor, was found to be elevated under hyperoxic CPB [22]. Therefore, vasoconstriction could be the reason for reduced FCD in our experiments. However, L/E cell interaction within capillary segments of the microcirculation can reduce FCD due to capillary plugging or endothelial cell swelling after a provocation like ischemia/ reperfusion. In our experiments hemodynamic parameters were stable throughout the entire experiment, thus tissue edema should not have developed due to mechanically severed perfusion. Compared to ischemia/reperfusion experiments using the skinfold chamber model of hamsters [20], the number of interacting white cells was lower in our setup, therefore we did not expect a significant effect caused by capillary plugging, which was never observed. ECC in our model seems to trigger the inflammatory response on a lower level, without rendering the tissue ischemic or even necrotic. The increase in RBC velocity in group 1 is in good agreement with results from Reber et al. [22], who found increased pulmonary shunting in patients undergoing hyperoxia during CBP. It is assumed that an increased RBC velocity under constant vessel diameters increases blood flow, consequently microvascular perfusion on the capillary level should increase. In our experiments FCD decreased under hyperoxia suggesting a significant shunt proximal to the capillary network.
Heparin in the dose of 2000 IU/kg body wt. has been shown to release high amounts of endothelium bound SOD into the systemic circulation in rodents [16]. In the case of the hamster baseline values of 60 -80 IU/ml can be increased to 500 IU/ml plasma by iv. injection of 2000 IU/kg body wt. heparin. Side effects like bleedings were not observed. The authors are aware that this fact makes it difficult to transfer individual drug doses and their actions among animal species or from animal species to humans. Yet our model allows to study pathophysiological and pharmacological mechanisms in respect to their specific individual actions in a standardized way.
The fact that FCD is reduced not only early after ECC but also 8 h after ECC supports the concept that hyperoxic ECC sets the stage for a self-amplificatory series of events, resulting in aggravation of tissue damage inflicted by ECC per se [15,19] as demonstrated by reduction of FCD in group I. Considering the half live of SOD (after heparin injection 50 min compared to 7 min when SOD was injected iv. [17]), recovery of FCD in group 2 at time point 8 h might be a result of inhibition of the above described events.
In our experiments inhibition of rolling was more pronounced compared to sticking, suggesting that heparin itself, via a selectin blocking mechanism, prevents rolling of leukocytes [18]. The time course of leukocyte activation, with a peak 8 h after ECC, suggests a delayed mechanism like generation and release of proteolytic enzymes and oxygen-derived radicals during and after ECC by activated neutrophils. This inflammatory response exacerbates under hyperoxia, leading to a significant increase of leukocyte activation compared to control. The fact that leukocyte adhesion is reduced not only early after ECC but also 8 h after ECC again supports the above mentioned concept of aggravation of tissue damage inflicted by ECC. Inasmuch as oxygen radicals are responsible to induce upregulation of adhesion molecules during ECC [4], the observed reduction in ECC-induced L/E cell adhesion may, at least in part, result from inhibition of the action of the oxidants in our experiments. Under hyperoxic ECC Heparin (SOD release) reduces the inflammatory response. However, not to a level as under ECC alone in the absence of hyperoxia, suggesting an overload of ROS under hyperoxia, thus overcharging the scavaging capacity of SOD.
A serious limitation of the study is the fact that no oxygenator, heater, cooler or cardiotomy suction devices are used is our experiments. Yet the standardized model of ECC in the awake hamster allows to investigate effects of blood contact with a foreign surface on the microcirculation under different levels of oxygenation. Reduction in FCD, increase in leukocyte accumulation and adherence to the microvascular endothelium of postcapillary venules and arterioles were major components of the inflammatory response to hyperoxia. Disturbances in the microcirculation under hyperoxia are supposed to result from the harmful action of oxygen during ECC in vivo. A better understanding of the microcirculation under hyperoxia may provide the basis for effective therapeutic interventions.