Selective Thromboxane A²Synthase Inhibition by OKY-046 Prevents Cardiopulmonary Dysfunction after Ovine Smoke Inhalation Injury

The current study was performed in 16 adult sheep of the Merino breed with a mean body weight of 41.2 ± 1.2 kg and a body surface area of 1.01 ± 0.02 m². All experiments were undertaken in compliance with the guidelines of the National Institutes of Health and the American Physiologic Society for the care and use of laboratory animals and were approved by the Animal Care and Use Committee of the University of Texas Medical Branch at Galveston.

After a 12-h fasting period (with access only to water), sheep were operatively instrumented for chronic study during deep isoflurane anesthesia. Catheters (18 gauge, 36 in; Parke Davis, Sandy, UT) were placed in the thoracic aorta and the inferior vena cava via  femoral arterial and venous cutdowns, respectively. In addition, a pulmonary artery catheter (Swan-Ganz, model 131F7; Baxter-Edwards Critical Care Division, Irvine, CA) was inserted into the right jugular vein through an 8.5-French percutaneous sheath introducer (Edwards Lifescience, Irvine, CA) and positioned in the pulmonary artery. For the cannulation of the caudal-mediastinal lymph node, a right-sided thoracotomy was performed at the level of the fifth intercostal space. The posterior end of the node was ligated by a modified technique originally described by Staub et al.  19In addition, a pneumatic occluder was placed around the inferior vena cava. Further, a left thoracotomy was performed at the level of the fifth intercostal space for implantation of pulse transit ultrasonic dimension transducers on the heart (LMT-53, 3-mm-diameter ceramic 5-MHz piezoelectric crystals; Crystal Biotech, Holliston, MA). Matched transducers were sutured to the anterior and posterior epicardium of the left ventricle for obtaining the maximum transverse external diameter in the plane of the minor axial circumference of the left ventricle. A silicon strain gauge pressure transducer (P5; Konigsberg Instrument, Pasadena, CA) was placed in the left ventricle via  a stab incision through the apex cordis. This apical transducer, in conjunction with a subdermal electrode, allowed continuous electrocardiographic monitoring of the heart. The left atrium was also cannulated with a silicon tube (Duralastic Silicone Tubing DT08, 0.062-in ID, 0.125-in OD; Allied Biomedical, Paso Robles, CA) for the measurement of left atrial pressure. Because we have previously demonstrated that systemic areas may significantly contaminate the lymph drainage,20the right and left hemidiaphragms were cauterized to reduce systemic contamination of the lung lymphatic duct. Then, the chest was closed, and the incision was infiltrated with lidocaine to aid perioperative recovery. All catheters exited the thorax at a site distant from the incision. An infusion of lactated Ringer’s solution was started at 2 ml · kg⁻¹· h⁻¹to assure sufficient postoperative hydration. During the subsequent recovery period of approximately 7 days, the animals were individually housed in metabolic cages and had free access to food and water. Animals were only included in the study if their hemodynamics, blood gases, lymph flow, leukocyte count, and core body temperature were within normal ranges. One day before the experiment began, all catheters were connected to pressure transducers (Baxter-Edwards Critical Care) with continuous flushing devices (heparin-saline solution, 3 U/ml). At the day of the experiment, each animal received a tracheostomy (Shiley, 10-mm ID, 13.3-mm OD; Tyco Healthcare Group LP, Pleasanton, CA) and a urinary bladder catheter (Dover, 12–14 French; Sherwood Medical, St. Louis, MO) using ketamine and halothane as anesthetics.

After a baseline measurement in the healthy state, sheep were randomly allocated to either the injured control group (control, n = 8) or the injured intervention group (OKY, n = 8) being treated with the specific thromboxane A²synthase inhibitor OKY-046 (Ozagrel, 3-[4-(1H-imidazol-1ylmethyl)phenyl]-2E-propanoic acid; Ono Pharmaceutical Co., Osaka, Japan). In the latter group, a 10-mg bolus infusion of OKY-046 was given 10 min before the injury and was followed by a continuous infusion of 10 μg · kg⁻¹· min⁻¹throughout the entire 24-h study period. The control group received only the vehicle (normal saline).

All sheep were subjected to smoke inhalation injury according to an established protocol.5–8,21Isoflurane (an ideal volatile anesthetic to maintain, but not to induce ovine anesthesia) was used for the operative procedures, whereas halothane (an inhalational anesthetic with a high analgesic potency) was administered during the induction of the injury. Ketamine was only used as a premedication to facilitate intubation. During the experiments, no anesthetics were administered, to avoid potential effects on the cardiopulmonary response. In brief, 4 × 12 breaths of cotton smoke (< 40°C) were delivered via  a modified bee smoker that was filled with 40 g burning cotton toweling and attached (via  a modified endotracheal tube containing an indwelling thermistor from a Swan-Ganz catheter) to the tracheostomy tube. Arterial carboxyhemoglobin plasma concentrations were determined immediately after each set of smoke inhalation and served as an index of acute lung injury.5,21After induction of the injury, anesthesia was discontinued, and the sheep were allowed to awaken.

Throughout the entire experiment, animals were equally mechanically ventilated (Servo Ventilator 900C; Siemens-Elema, Stockholm, Sweden). Because sheep have a higher pulmonary compliance than humans, they were ventilated with a tidal volume of 15 ml/kg. To prevent atelectasis, a positive end-expiratory pressure of 5 cm H²O was applied. During the first 3 h after injury, the inspiratory oxygen fraction was set at 100%, and the respiratory rate was set at 30 breaths/min to induce rapid carbon monoxide clearance. Ventilation was then adjusted to maintain sufficient oxygenation (arterial oxygen saturation > 95%, arterial oxygen tension > 90 mmHg) whenever possible.

All sheep were studied in the standing position while unanesthetized and undistressed. During this time, they were fluid resuscitated with lactated Ringer’s solution (3 ml · kg⁻¹· h⁻¹) and had free access to food from 1 h after injury.

The ultrasonic dimension transducers were connected to a Valpey-Fischer VF-1 Dimension System (Valpey-Fischer, Hopkinton, MA). The ultrasonic signal was amplified and continuously monitored using an oscilloscope (model 2213; Tektronix Inc., Beaverton, OR). Mean arterial pressure, mean pulmonary arterial pressure, and left atrial pressure were measured with Statham-Gould P23XL-1 transducers. The pressure transducers and the electrocardiogram were connected to a physiologic recorder (model OM9; Electronics for Medicine, Pleasantville, NY). The highest value of the left ventricular pressure was assumed to be the same as the peak aortic pressure; the lowest value was assumed to be 0 pressure (atmospheric pressure as reference). At the end of the study, these values were again checked, and it was certified that there was no drift over 24 h in each experiment. All dimensions, pressures, and electrocardiographic data were digitalized at 250 Hz with an RTI-800 analog to digital converter (Analog Devices Inc., Norwood, MA) and stored on magnetic media. Data were collected and averaged over one respiratory cycle, and at least five cardiac cycles were normalized to exclude the effect of respiration on the left ventricular dimension and pressure. Cardiac output was measured using the thermodilution technique with a Swan-Ganz catheter connected to a cardiac output computer (model 9520; Edwards Laboratory, Irvine, CA). Cardiac index, systemic vascular resistance index, and pulmonary vascular resistance index were calculated using standard equations. Lung lymph flow was measured with graduated test tubes and a stopwatch. Lymph and plasma protein concentrations were determined with a protometer (Refractometer; National Instruments, Baltimore, MD). Protein flux was calculated by multiplying lung lymph flow by the lymph to plasma protein concentration ratio. Heparinized samples of arterial plasma or lymph, collected with 2 μg/ml 1-phenyl-3-pyrazolidone, were immediately frozen at −80°C after collection and centrifugation. Subsequently, thromboxane B²and 6-keto-prostaglandin F^1αwere measured by radioimmunoassay.22The delivery of thromboxane B²to the lung was calculated by multiplying lung lymph flow with the concentration of thromboxane B²in the lung lymph. In the same way, the delivery of 6-keto-prostaglandin F^1αto the lung was calculated. Plasma conjugated dienes, lipid peroxidation products, were measured spectrophotometrically (Spectronic 1001; Milton Roy Co., Houston, TX) at a wavelength of 233 nm.23Oxygen and carbon dioxide tensions in arterial and mixed venous blood were analyzed at 37°C and were corrected for core body temperature (System 1302; Instrumentation Laboratory, Lexington, MA). Oxygen saturation and carboxyhemoglobin concentrations were determined with an oximeter (CO-Oximeter 682; Instrumentation Laboratory).

An intraventricular pressure transducer and crystalographic dimension transducers were implanted for the determination of the end-systolic pressure–diameter relation. This relation is linear, unaltered by changes in preload or afterload, and sensitive to alterations in the inotropic state of the left ventricle.24The external minor axis was used in this analysis because it correlates closely with simultaneous left ventricular volumes under all conditions, during both diastole and systole.25End systole was defined as the point of maximum elasticity calculated by the ratio of instantaneous left ventricular pressure to instantaneous left ventricular diameter. In each group, the inferior vena cava was briefly occluded to reduce the preload of the left ventricle during determinations of the end-systolic pressure–diameter relation. Simultaneous left ventricular pressure and diameter loops were used to determine the end-systolic pressure–diameter relation of the left ventricle. The inferior vena cava occlusion was performed to minimize reflex sympathetic activity to the heart while measuring end-systolic pressure–diameter points for several curves. Without this approach, the end-systolic pressure–diameter maximum might have been changed. A line of best fit was described by the computer by the means of linear regression. The computer calculated the slope of the linear regression line and reported it as end-systolic pressure–diameter relation. Notably, only measurements with an r  value greater than 0.9 were accepted.

After completion of the 24-h experiment, the animals were anesthetized with ketamine (15 mg/kg) and killed by intravenous injection of 60 ml saturated potassium chloride. Immediately after death, the proper positions of the ultrasonic dimension transducers were verified in every sheep.

Data are presented as mean ± SEM. For statistical analysis, Sigma Stat 2.03 software (SPSS Inc., Chicago, IL) was used. Using a Kolmogorov-Smirnov test for goodness of fit to normal distribution, normality was obtained for all measurements. A two-way analysis of variance for repeated measurements with appropriate Student-Newman-Keuls post hoc  comparisons was used to calculate differences within and between groups. A P  value of less than 0.05 was considered as statistically significant. The correlation coefficient r  was determined using the Pearson product moment formula.

The arterial carboxyhemoglobin plasma concentrations determined immediately after the fourth set of cotton smoke exposure averaged 65 ± 5% in the control group and 67 ± 6% in sheep treated with OKY-046, indicating that both groups were equally injured. With aggressive fluid resuscitation, all animals survived the 24-h study period.

Throughout the entire experiment, lung lymph flow was significantly higher in the control group as compared to the OKY group (fig. 1). Because there were no changes in the lymph to plasma protein concentration ratio in either group, the change in protein flux was similar to the changes in lung lymph flow (table 1).

There was a marked increase in thromboxane B²delivery to the lung in injured control animals, whereas pulmonary delivery of thromboxane B²remained unchanged in the OKY group (fig. 2A). Similarly, pulmonary delivery of 6-keto-prostaglandin F^1αprogressively increased in the control group, but was unaffected in the OKY group from baseline to 12 h after injury (fig. 2B). Thromboxane B²plasma concentrations remained constant in either group, whereas there was an immediate and sustained increase in 6-keto-prostaglandin F^1αplasma concentration within each group but not between the two groups (fig. 3).

Hemodynamics of the systemic and pulmonary circulation are summarized in table 2and show that the smoke-associated depression in cardiac index was attenuated and delayed in the OKY group. During the entire observation period mean arterial pressure, left atrial pressure and heart rate remained unchanged in either group. In the control group, however, there was a progressive increase in systemic vascular resistance index as compared to baseline. At 24 h after injury, systemic vascular resistance index was significantly higher in the control group than in the OKY group. Similarly, we noted successive increases in mean pulmonary arterial pressure and pulmonary vascular resistance index in the control group that were significantly higher than in the OKY group (table 2).

The changes in left ventricular function are depicted in figure 4. Smoke inhalation injury contributed to an immediate and sustained reduction in the end-systolic pressure–diameter relation (control), whereas the end-systolic pressure–diameter relation remained unchanged in the intervention group (OKY). From 8 to 24 h after injury, the end-systolic pressure–diameter relation was significantly lower in the control group than in the OKY group. Likewise, left ventricular +dp/dt remained unchanged in the OKY group. In the control group, maximum +dp/dt was significantly depressed and lower than in the OKY group from 12 to 24 h after injury.

Compared with baseline, plasma conjugated dienes were significantly increased in the control group at 12 and 24 h after smoke exposure and were higher than in the OKY group. During the entire experiment, there were no changes in plasma conjugated dienes in the OKY group (fig. 5).

OKY infusion also attenuated the progressive increase in airway resistance observed in the control group, as indicated by decreases in peak and pause airway pressures (data not shown).

Using an established and clinically relevant ovine model of smoke inhalation injury,5–8,21we investigated the role of thromboxane A²in the pathogenesis of cardiopulmonary dysfunction. The intervention group was treated with OKY-046, a 1-alkyl imidazole derivative with an IC⁵⁰of 11 nm, acting as a specific thromboxane A²synthase inhibitor.26,27The major finding was that cotton smoke exposure increased lung lymph content of thromboxane B²(stable end product of thromboxane A²), which in turn augmented transvascular fluid flux and contributed to significant decreases in both end-systolic pressure–diameter relation and maximum +dp/dt. Infusion of OKY-046 attenuated not only the increased pulmonary microvascular permeability seen after smoke inhalation injury, but also prevented myocardial depression.

The complex pathophysiology of smoke inhalation injury involves multiple factors, some well known and others still undetermined. In the past, we have demonstrated that cotton smoke exposure results in significant epithelial cell injury of the bronchial tree, which in turn contributes to the release of cytotoxic mediators, such as proteolytic enzymes, oxygen free radicals, and eicosanoids.6,28The latter are synthesized by phospholipase A², releasing arachidonic acid from membrane phospholipids. Subsequently, cyclooxygenase converts arachidonic acid to the precursor prostaglandin G². Prostacyclin and thromboxane A²are downstream metabolites produced by prostacyclin synthase and thromboxane synthase, respectively. Prostacyclin is an abundant product of endothelial cells that causes vasodilation, whereas thromboxane A²promotes significant vasoconstriction.29Because recent studies in prostacyclin and thromboxane A²genetically deficient mice confirmed that the balance between those two eicosanoids modulates the cardiovascular response to injury,30we measured their stable end products, i.e. , prostaglandin F^1αand thromboxane B².

The cardiopulmonary derangements in the untreated control group was linked to increased thromboxane B²delivery to the lung. Despite an augmented release of prostacyclin via  the cyclooxygenase pathway, pulmonary vascular resistance index and mean pulmonary arterial pressure increased significantly over time. This finding is in full agreement with the study of Redl et al. ,14showing that increased prostaglandin F^1αplasma concentrations in endotoxemic sheep was ineffective to compensate for increased thromboxane B²concentrations and did not prevent pulmonary hypertension. Infusion of OKY-046, the specific thromboxane A²synthase inhibitor, attenuated the pulmonary vasoconstrictive response and improved right ventricular function, as indexed by improved ejection fraction and increased cardiac output. In the current study, infusion of OKY-046 prevented not only the increase in thromboxane B²but also in mean pulmonary arterial pressure and pulmonary vascular resistance index. This is of special clinical importance, because increased pulmonary resistance may lead to cardiac dysfunction/failure.31The fact that thromboxane B²was only increased in the lung lymph but not in the plasma suggests that pulmonary inflammation triggers thromboxane synthesis in the airway. This postulation is supported by the finding of Takimoto et al.  32showing that endothelin 1, secreted from human bronchial epithelial cells, evokes the release of thromboxane B²in an autocrine manner. Similar results have also been reported by Janssens et al.  33Using a sheep model of acrolein smoke–induced acute lung injury, the authors demonstrated that locally increased thromboxane B²concentrations are implicated in the pathogenesis of pulmonary hypertension.

Reactive oxygen species are released in various forms of lung injury.34–36In accord with previous studies,23we noted significantly increased plasma concentrations of conjugated dienes in response to the injury. Because conjugated dienes are lipid peroxidation products, their presence in the plasma is a direct evidence of an interaction of oxygen free radicals with lipid materials in the sheep. Because inhibition of thromboxane A²synthesis prevented the formation of conjugated dienes, it is most likely that thromboxane A²stimulated reactive oxygen species formation. This hypothesis is supported by the study of Paterson et al. ,37demonstrating that there is a direct relation between the plasma concentrations of thromboxane B²5 min after reperfusion and the magnitude of the subsequent intracellular H²O²production. Furthermore, these authors reported that activated polymorphonuclear cells, in turn, lead to more thromboxane synthesis, a positive feedback mechanism resulting in further neutrophil activation. In this regard, it is especially important that neutrophils represent a main source of thromboxane release after acute lung injury.38Moreover, it must be considered that polymorphonuclear cells play a major role in the development of the primary lung injury associated with ovine smoke inhalation injury and may contribute to its systemic manifestation.39Therefore, it is most likely that OKY-046 limited neutrophil accumulation in the lung, an assumption that is supported by the study of Goldman et al.  40Using a rat model of acid aspiration, the authors demonstrated that OKY-046 limited adhesion of polymorphonuclear cells.

The most significant finding of the current study was that the intrinsic myocardial contractility was not depressed to the same extent in the treated animals as in the control group. With inhalation injury, maximum elastance of left ventricle end-systolic pressure–volume relations decreased by more than 50%. When the question arises as to whether carbon monoxide exposure, per se , may have accounted for the decrease in myocardial contractility, it is noteworthy that we have previously demonstrated that smoke-associated myocardial depression is not linked to carbon monoxide intoxication; the exact mechanism, however, remained undetermined.8When viewing the above-discussed studies together with the results of the present investigation, it seems that thromboxane B²–mediated myocardial dysfunction was triggered by neutrophil accumulation in the injured lung. This hypothesis is supported by the study of Evangelista et al. ,38reporting that acute lung inflammation is associated with infiltration of polymorphonuclear cells and contributes to thromboxane A²–mediated myocardial ischemia in a rabbit model.

Ovine smoke inhalation injury was also associated with a marked increase in lung lymph flow, similarly to what has been previously observed in our model.39We noted that administration of OKY-046 markedly blunted this response. Likewise, Enkhbaatar et al.  41demonstrated that the cyclooxygenase inhibitor ketorolac reduces pulmonary edema after combined burn and smoke inhalation injury, in part via  inhibiting nitric oxide synthesis. These findings suggest that eicosanoid overproduction and the nitric oxide pathway are linked in the pathophysiology of smoke inhalation injury. However, future studies are needed to address this issue in greater detail.

This study has some limitations that we want to acknowledge. First, we cannot assure that exogenous administration of OKY-046 produces identical responses in sheep and patients when used under similar conditions. Because we did not investigate regional blood flow, it also remains undetermined how thromboxane A²synthase inhibition affected gastrointestinal blood flow in the current study. However, Sakurai et al.  42previously reported an early reduction in mesenteric blood flow in sheep with burn and smoke inhalation injury that corresponded to the period of reduced systemic blood flow. Because OKY-046 prevented the increase in systemic vascular resistance index and alleviated the decrease in cardiac index, it is likely that mesenteric blood was also improved by thromboxane inhibition. This notion is reinforced by the study of Tokyay et al. ,43demonstrating that OKY-046 improved postburn mesenteric blood flow and decreased the rate of bacterial translocation in a porcine model of burn injury.

Although OKY-046 may prove therapeutic, we want to emphasize that the timing of administration may be critical, especially given the often disappointing clinical responses to other inhibitors in past trials in the critically ill. In this regard, it is also noteworthy that Ishitsuka et al.  18reported that a pretreatment with OKY-046 was more effective than a posttreatment of guinea pigs with acute respiratory distress syndrome. Therefore, future studies are needed to test the efficacy of this specific thromboxane A²inhibitor when it is given after smoke inhalation injury.

Notably, we have already demonstrated that ovine smoke inhalation injury is not only associated with increased thromboxane production, but also with excessive nitric oxide synthesis.44Because inhibition of each product only attenuated cardiopulmonary dysfunction but did not block the underlying pathogenesis, future studies are needed to determine whether a combination therapy of selective nitric oxide and thromboxane A²synthase inhibitors are useful to further reduce the degree of injury.

In this prospective study, we showed that thromboxane A²release secondary to smoke inhalation contributes to myocardial depression and increased pulmonary microvascular fluid flux. Because OKY-046 markedly attenuated cardiopulmonary derangements and lipid peroxidation, specific thromboxane A²synthase inhibition seems to be a promising, goal-directed therapeutic target in the setting of smoke inhalation injury. Future studies investigating the role of thromboxane A²in the setting of acute respiratory distress syndrome are warranted and may help to test preventative measures and therapeutic strategies.

The authors thank Nettie Biondo (Research Associate, University of Texas Medical Branch at Galveston, Galveston, Texas) and John R. Salsbury, M.Sc. (Research Associate, University of Texas Medical Branch at Galveston), for expert technical assistance.