INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis is a common cause of the adult respiratory distress syndrome (ARDS). Approximately 40% of patients with severe sepsis develop ARDS while over 90% of sepsis patients develop clinically significant hypoxemia. It has been previously suggested that the presence of pulmonary hypertension indicates a poor prognosis for patients with ARDS associated with sepsis (1). The etiology of the pulmonary hypertension associated with ARDS is not completely characterized, but it does not appear to be strictly related to hypoxemia, acidosis, or left heart failure. Pulmonary hypertension could be more than a poor prognostic marker in ARDS and may directly contribute to the pathogenesis of the ARDS associated lung injury. If increases in pulmonary arterial pressure are transmitted to the pulmonary microvasculature, a marked increase in fluid and solute transudation might occur from the capillaries to the pulmonary interstitial space across an abnormally permeable microvasculature. The experiments described in this article were designed in an attempt to elucidate the mechanism responsible for the sustained pulmonary hypertension observed following endotoxemia in a sheep model of ARDS.

The infusion of low doses (0.5 µg/kg) of endotoxin (LPS) into chronically instrumented awake sheep causes a well-characterized sequence of pathophysiologic abnormalities, including marked changes in pulmonary hemodynamics, lung mechanics, lung fluid and solute exchange, and hypoxemia (4). The response to 0.5 µg/kg LPS is reproducible when given on two separate occasions separated by five or more days (4). Higher doses of LPS (4 µg/kg) can cause frank pulmonary edema, respiratory failure and death (5). Despite the fact that low and high doses of LPS in sheep do not cause systemic hypotension and the type of hemodynamic changes observed with clinical sepsis in humans, results from the low dose LPS model in sheep have proven relevant to both ARDS and sepsis in humans (4). Interventions which have not proven efficacious in the sheep model have also failed to have beneficial effects in humans and similar positive results have been observed in humans to those obtained with the same interventions in the LPS sheep model.

It has proven useful to separate the pathophysiologic effects of LPS into early (changes occurring within the first hour of LPS infusion) and late alterations (changes occurring between 2 and 5 h after LPS). The infusion of 0.5 µg/kg of Escherichia coli endotoxin into sheep results in early severe pulmonary hypertension, hypoxemia, and altered lung mechanics. These early abnormalities are followed, from 2-5 h after LPS, by moderate persistent pulmonary hypertension, less severe abnormalities in lung mechanics, and increased pulmonary microvascular permeability. These late changes in lung fluid and solute exchange are felt to be analogous to ARDS in humans (4). The infusion of 0.5 µg/kg of LPS does not cause marked changes in cardiac output or shock (4).

Previous experimental evidence suggests that cyclooxygenase products of arachidonic acid metabolism mediate the early changes in lung mechanics and pulmonary artery pressure observed after endotoxemia in sheep. Studies using cyclooxygenase inhibitors (6), thromboxane synthase inhibitors (9), and thromboxane-prostaglandin endoperoxide receptor antagonists (14, 15) suggest that thromboxane A₂ is responsible for the early changes in pulmonary hemodynamics and lung mechanics. Cyclooxygenase inhibitors do not attenuate the late pulmonary hypertension nor the late change in lung fluid and solute exchange.

Endotoxemia causes increased plasma endothelin-like immunoreactivity in animals (16), and plasma endothelin (ET) concentrations have been shown to correlate with the severity of illness in patients with sepsis (20). Endothelin-1 (ET-1) and endothelin-3 (ET-3) cause contraction of isolated pulmonary arterial and venous rings (21, 22) and are potent pulmonary vasoconstrictors in several animal species (23). We hypothesize that endothelin is responsible for the late non-thromboxane-dependent pulmonary hemodynamic changes seen in endotoxemia.

To test this hypothesis, we employed BMS182874. BMS182874 in vitro selectivity inhibits a subtype of ET-receptor called the ET_A receptor. There appear to be at least two different subtypes of ET-receptors, ET_A and ET_B, which are functionally classified by their varying affinity for the three endothelins (26, 27). ET_A has a high affinity for ET-1 and ET-2 but not for ET-3, while ET_B is nonselective for the three endothelins. ET_A and ET_B receptors have been cloned, and Nakamichi and colleagues have shown that porcine pulmonary ET_A receptors are preferentially localized in the bronchi and pulmonary vasculature, whereas ET_B receptors are more abundant in the parenchyma (28). Since little data are available on the functional distinction between ET_A and ET_B receptors in vivo and no information was available on the selectivity of BMS182874 in vivo in sheep, we investigated the ability of BMS182874 to block the actions of ET-1, ET-2, and ET-3. Based upon results obtained during the BMS182874 and LPS experiments, we further studied the effects of BMS182874 on dose-response curves to U46619. U46619 is a prostaglandin H₂-analogue which selectively acts through interaction with the thromboxane-prostaglandin endoperoxide receptor. These studies allowed us to determine if BMS182874 functioned in vivo as a thromboxane-receptor antagonist in vivo in addition to its actions as an ET-receptor antagonist. We studied the effects of BMS182874 on LPS-induced change in hemodynamics, lung mechanics, and lung fluid and solute exchange in chronically instrumented awake sheep. Thromboxane B₂ concentrations were measured in blood and lung lymph following LPS. The thromboxane data allowed us to rule out BMS182874 as having the pharmacologic ability to inhibit cyclooxygenase, the enzyme system responsible for the production of thromboxane, the prostaglandins, and prostacyclin.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sheep Preparation

Yearling sheep (25-35 kg) of both sexes were prepared surgically as previously described (29). Through a right thoracotomy, the caudal mediastinal lymph node was exposed and the efferent lymph vessel was cannulated with Silastic tubing (internal diameter 0.12 inches). The tubing was then externalized. Through a second right thoracotomy, the caudal portion of the lymph node was ligated to interrupt the nonpulmonary lymphatics. Lymph obtained in this fashion is predominantly of pulmonary origin (30). Two Silastic envelopes attached to Silastic tubing were placed in the pleural cavity, and the tubing was externalized to measure pleural pressure. Through a left thoracotomy, Silastic catheters were inserted directly into the main pulmonary artery and the left atrium. A Transonic (Ithaca, NY) 16S ultrasonic flow probe was positioned around the main pulmonary artery. Vascular catheters were also inserted into the thoracic aorta and superior vena cava through cervical vessels. On a different occasion, a tracheostomy was performed. The animals were allowed to recover for several days prior to investigation. A size 10 Shiley cuffed tracheostomy tube (Shiley, Irvine, CA) was inserted at the time of experimentation.

Lung Mechanics

Sheep were studied unanesthetized while standing in a whole-body pressure-compensated integrated-flow plethysmograph. Ropes were used as passive restraints to prevent the sheep from lying down during the experiment. The sheep's tracheostomy tube was connected to an external valve via noncollapsible flexible tubing. A constant bias flow of room air was used to reduce the effective dead space of the tubing. The sheep were studied, at all times, on room air. Tidal volume (V) was measured by pressure compensating the integrated signal from the plethysmographic pressure transducer. Flow was obtained by electronically differentiating the volume signal. Pleural pressure (P_PL) was measured directly from the Silastic catheter and envelope in the pleural space. Airway opening pressure (P_AO) was measured by a multiple-side-hole catheter positioned 0.5 cm past the distal end of the tracheostomy tube. Transpulmonary pressure (P_TP) was the pressure difference between P_PL and P_AO. All pressure signals were measured using Validyne MP-45 pressure transducers and amplification equipment (Validyne Engineering, Northridge, CA). The signals from the pressure transducers, catheters, and Silastic envelopes were tuned to eliminate phasic distortion to 20 Hz. These methods have been described previously in detail (8).

Before each measurement of lung mechanics, the sheep's lungs were inflated to 40 cm H₂O airway opening pressure using the bias flow and an occluded airway. Simultaneous V/flow and V/P_TP curves were then recorded during spontaneous respiration on a Tektronix dual-beam storage oscilloscope (Tektronix, Beaverton, OR) and photographed for calculation of dynamic compliance (C_dyn) and resistance to air flow across the lung (R_L). C_dyn was calculated as V divided by P_TP at points of zero flow and expressed in liters per centimeter H₂O at body temperature pressure standard (BTPS). R_L was calculated by use of the method of von Neergaard and Wirz (31) by dividing P_TP by flow at mid-tidal volume and was expressed as centimeters H₂O per liter per second at BTPS.

Thromboxane B₂ (TxB₂) Measurements

TxB₂ was measured in plasma and lung lymph specimens collected at 30-min intervals throughout the experiment. Analyses were performed by radioimmunoassay using rabbit anti-TxB₂ antibodies obtained from Dr. J. Bryan Smith (Cardeza Foundation, Philadelphia, PA). The anti-TxB₂ antibody cross reacts < 3% with PGD₂, and < 1% with PGE₂, PGF₂, and 6-keto-PGF₂. Authentic TxB₂ was purchased from Cayman Chemical Company (Ann Arbor, MI). Radiolabeled TxB₂ was purchased from New England Nuclear (Boston, MA). Bovine gamma-globulins, Trizma buffer, and ammonium sulfate were purchased from Sigma Chemical Company (St. Louis, MO).

The radiolabeled ligand (approximately 2,000 cpm/tube) was mixed with bovine gamma-globulins (10 mg/ml in Trizma, pH 7.4). 100 microliter aliquots of this mixture were combined with 100 microliter aliquots of sample or appropriate standard dilutions. The binding reaction was initiated by addition of 100 microliters of the antibody diluted to yield 60% binding of the label. The binding reaction continued for 60 min at 37° C and was terminated by precipitation of the immune complexes with ammonium sulfate at a final concentration 50% of saturation. After centrifugation at 2,500 g at 4° C for 10 min, 300 microliters of supernatant were counted in Aquasol (New England Nuclear). Each sample was assayed in duplicate, and duplicate determinations differed by < 10%. The detection limit of this assay was  20 pg. If TxB₂ determinations exceeded by greater than two standard deviations of the mean of all other determinations for a given time point following LPS or drug, the entire data set of TxB₂ measurements from that animal was excluded from statistical analysis.

Hemodynamics

We continuously recorded mean pulmonary artery pressure (P_PA), mean left atrial pressure (P_LA), and mean aortic pressure (P_SA) using fluid-filled pressure transducers (model 1208C; Hewlett-Packard Co., Palo Alto, CA). The level of the left atrium was the zero reference for all vascular pressure measurements. Cardiac output was continuously monitored using a Transonic System Inc. (Ithaca, NY) T101 Ultrasonic Blood Flow Meter and the 16S ultrasonic probe.

Experimental Protocol

The effects of BMS182874 (supplied by Suzanne Moreland at Bristol-Myers Squibb, Princeton, NJ) on the response to LPS were studied in six chronically instrumented unanesthetized sheep. Each sheep served as its own control and was studied three times on room air: once with the vehicle for BMS182874 (5% sodium bicarbonate) and LPS, once with BMS182874 alone, and once with LPS after BMS182874 pre-treatment. The order of experimentation was varied to exclude sequential bias. A minimum of 5 d was allowed between experiments (LPS, controls, exogenously administered ETs, or U466619).

At the beginning of each experiment, the physiological measurements were recorded until a stable baseline was obtained for at least 1 h. BMS182874 was prepared by dissolving 120 mg/kg BMS182874 in 252 ml 5% sodium bicarbonate. After baseline measurements, infusion of BMS182874 (or an equal amount of vehicle) via the central venous line was started as a loading dose of 50 mg/kg over 30 min followed by a continuous infusion of 10 mg/kg/h through the remainder of the experiment. One hour after beginning the drug (or vehicle) infusion, 0.5 µg/kg E. coli endotoxin in sterile 0.9% sodium chloride (lipopolysaccharide E. coli 055:B5; Difco Laboratories, Detroit, MI) was infused over 15 min via the pulmonary artery catheter (total volume less than 30 ml). Measurements were made for 5 h after the start of the LPS infusion. Sheep received no additional fluid management.

Throughout the experiment, mean intravascular pressures and cardiac output were continuously recorded and averaged for intervals of 15 min. Lung mechanics were measured every 15 min throughout the experiment. Lung lymph flow was measured every 15 min and collected every 30 min for analysis. Arterial blood was collected every 30 min for white blood cell count and blood gas determination. Both the blood and lymph samples were collected in the presence of EDTA and indomethacin. The blood was centrifuged (760 g for 15 min) after initial analysis and the resultant plasma stored at 20° F for later measurements of plasma protein concentration and thromboxane B₂ levels.

To help define the ability of BMS182874 to block the effects of exogenously administered endothelin, we administered ET-1, ET-2, and ET-3 (Calbiochem-Novabiochem Corporation, La Jolla, CA) by intravenous bolus before and after treatment with BMS182874. P_PA and C_dyn were allowed to return to baseline values between ET doses (a minimum of 30 min between doses). After each ET bolus, P_PA was monitored continuously and C_dyn was calculated at 1-min intervals until both variables returned to baseline. These studies were used to determine the BMS182874 dose used for the LPS studies (50 mg/kg loading dose followed by 10 mg/kg/h maintenance infusion). To confirm the effectiveness of BMS182874 in the sheep treated with LPS, bolus injections of 30 µg ET-1 were given to sheep via the pulmonary artery at the end of their six-hour infusion of BMS182874.

In an attempt to clarify the mechanism of action of BMS182874, U46619 (Caymen Chemical) a prostaglandin H₂-analogue, which mimics the action of thromboxane A₂ (32), was administered to two sheep with and without pretreatment with BMS182874 (as outlined above) while monitoring P_PA and P_SA. U46619 (dissolved in 0.9% sodium chloride) was administered by constant intravenous infusion at incremental doses (0.01, 0.05, 0.10, 0.25, 0.50, 1.00, 2.50, 5.0, and 10.0 µg/kg/min), increasing the dose every eight minutes, thus providing sufficient time to achieve a new equilibrium pressure for both P_PA and P_SA. After an initial U46619 versus P_PA dose response curve was established and pressures returned to baseline BMS182874 was administered as outlined above and the U46619 dosing was repeated. When two dose response curves are done on the same day with U46619 alone, reproducible changes in P_PA and P_SA are observed.

Other Methods

White blood cells were counted using a Coulter Counter (model ZBI; Fear Electronics, Hialeah, FL). Total protein concentrations in plasma and lung lymph were measured with an automated system (AutoAnalyzer; Technicon Instruments Corp., Tarrytown, NY) by a modified biuret method (33). Lymph protein clearance (C_LP) was calculated by multiplying the lymph flow by the lymph-to-plasma protein concentration ratio (L/P). Arterial gas tensions for oxygen and carbon dioxide and pH were determined by a Corning blood gas analyzer (238 pH/Blood Gas Analyzer; Corning Medical and Scientific, Medfield, MA). The alveolar-to-arterial oxygen difference (AaPO₂) on room air was calculated using the alveolar gas equation with a fixed respiratory exchange ratio of 0.8.

Statistics

The data were analyzed using parametric repeated measures analyses of variance with multiple comparisons by Newman-Keuls test (Statistica/W; StatSoft, Tulsa, OK). Log₁₀ transformation of the P_PA, P_SA, and Q_L data was done to ensure normal distribution and/or homogeneity of variance. A value of p < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ET-1, ET-2, ET-3 and BMS182874

Bolus injections into the pulmonary artery of ET-1, ET-2, or ET-3 caused reproducible changes in pulmonary hemodynamics and lung mechanics. The changes were similar in pattern and magnitude for all three ETs and consisted of an acute substantial increase in P_PA and decrease in C_dyn within two minutes of administration. These effects were transient, with P_PA and C_dyn returning to baseline within 5 and 15 min, respectively. The threshold for a response was between 1 and 3 µg with increasing responses noted with doses of 10 and 30 µg bolus injection of any of the ETs. A dose of 100 µg caused death. Thirty micrograms of ET-1 caused P_PA to increase from 18.6 ± 1.0 to 47.6 ± 2.8 cm H₂O. Thirty micrograms of ET-2 caused P_PA to increase from 19.0 ± 1.7 to 42.8 ± 4.8 cm H₂O. 30 µg of ET-3 caused P_PA to increase from 17.8 ± 1.4 to 48.6 ± 5.4 cm H₂O. ET-1, ET-2, and ET-3 caused C_dyn to decrease to 39.7 ± 4.4, 45.5 ± 5.8, and 48.4 ± 6.7% of baseline, respectively. BMS182874 administered as a 50 mg/kg loading dose followed by 10 mg/kg/h maintenance infusion entirely blocked the effects of 30 µg bolus injection of ET-1, ET-2, and ET-3.

Vehicle and LPS

The order in which the experiments were performed had no effect on the changes in physiologic variables measured. Vehicle (5% sodium bicarbonate) infusion caused no significant changes in any of the variables measured (including blood pH). LPS infusion caused marked changes in pulmonary hemodynamics, lung mechanics, and lung fluid and solute exchange which were similar to those previously described (3, 4, 9, 12, 13, 20, 21). As shown in Figure 1, P_PA rapidly increased from a baseline of 17 ± 1 to a peak of P_PA 55 ± 5 cm H₂O (mean ± SEM) at approximately 45 min after initiation of the LPS infusion. P_PA then declined but remained significantly greater than both the pre-LPS baseline (p < 0.05) and the time-matched BMS182874 alone values (p < 0.05) throughout the experiment. P_SA did not significantly change as a result of LPS infusion (Figure 2). P_LA and cardiac output did not change significantly following vehicle and LPS.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Effect of BMS182874 on LPS-induced changes in pulmonary artery pressure. Data are expressed as the mean ± SEM. The 2-h point is the peak value. The 90-min point represents the mean of the data over the second hour. Open circles represent LPS alone experiments. Solid circles represent LPS plus BMS182874 experiments. Solid squares represent BMS182874 alone experiments. n = 6. An asterisk indicates that the data point is significantly different (p < 0.05) from baseline and a diamond indicates that the data point is significantly different (p < 0.05) from the time-matched LPS plus BMS182874 data point.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Effect of BMS182874 on LPS-induced changes in systemic arterial pressure. Data are expressed as the mean ± SEM. Open circles represent LPS alone experiments. Solid circles represent LPS plus BMS182874 experiments. Solid squares represent BMS182874 alone experiments. n = 6. An asterisk indicates that the data point is significantly different (p < 0.05) from baseline, and a triangle indicates that the data point is significantly different (p < 0.05) from the time-matched LPS alone data point. PSA in the LPS plus BMS182874 was not significantly different from the BMS182874 alone data.

Q_L increased markedly during the early pulmonary hypertension from 1.9 ± 0.3 ml/15 min to a maximum of 5.1 ± 1.2 ml/15 min at 2 h after beginning LPS and remained significantly elevated throughout the remainder of the experiment (Figure 3). Data for L/P and C_LP are presented in Figures 4 and 5, respectively. There was considerable variability in the Q_L and LPS data and the changes were not statistically significant but are qualitatively similar to those observed in earlier studies (6).

View larger version (13K): [in this window] [in a new window]   Figure 3.   Effect of BMS182874 on LPS-induced changes in lung lymph flow. Data are expressed as the mean ± SEM. Open circles represent data from LPS alone experiments. Solid circles are data from LPS plus BMS182874 experiments. Solid squares are data from BMS182874 alone experiments. n = 6. An asterisk indicates that the data point is significantly different (p < 0.05) from baseline, and a diamond indicates that the data point is significantly different (p < 0.05) from the time-matched LPS plus BMS182874 data point.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Effect of BMS182874 on LPS-induced changes in lymph/ plasma protein ratio. Data are expressed as the mean ± SEM. Open circles represent LPS alone experiments. Solid circles are data from LPS plus BMS182874 experiments. Solid squares are data from BMS182874 alone experiments. n = 6.

View larger version (13K): [in this window] [in a new window]   Figure 5.   Effect of BMS182874 on LPS-induced changes in lymph protein clearance. Data are expressed as the mean ± SEM. Open circles represent data from LPS alone experiments. Solid circles are data from LPS plus BMS182874 experiments. Solid squares are data from BMS182874 alone experiments. n = 6.

Plasma thromboxane B₂ levels acutely increased within one hour after LPS from a baseline of 190 ± 57 to 1,640 ± 551 pg/ ml and then returned to baseline (Figure 6). Lung lymph thromboxane B₂ similarly increased from 289 ± 89 to 4,032 ± 2,456 pg/ml with a similar time course (Figure 7). These changes did not reach statistical significance but are similar to those previously reported (20).

View larger version (13K): [in this window] [in a new window]   Figure 6.   Effect of BMS182874 on LPS-induced changes in plasma thromboxane B2 levels. Data are expressed as the mean ± SEM. Open circles represent data from LPS alone experiments. Solid circles are data from LPS plus BMS182874 experiments. Solid squares are data from BMS182874 alone experiments. n = 4.

View larger version (12K): [in this window] [in a new window]   Figure 7.   Effect of BMS182874 on LPS-induced changes in lung lymph thromboxane B2 levels. Data are expressed as the mean ± SEM. Open circles represent data from LPS alone experiments. Solid circles are data from LPS plus BMS182874 experiments. Solid squares are data from BMS182874 alone experiments. n = 4.

As shown in Figure 8, after LPS infusion C_dyn declined acutely to a nadir of 43 ± 9% of pre-LPS baseline at approximately 45 min after beginning LPS and then quickly trended toward baseline. Concomitant with the decline in C_dyn, AaPO₂, increased to 151 ± 14% of baseline at 1 h after LPS and remained significantly elevated at 5 h after LPS (Figure 9). Blood pH did not change significantly throughout the study period. WBC declined from a baseline 9,937 ± 735 to a nadir of 3,294 ± 687 cells/mm³ 45-75 min after starting LPS infusion and then increased back to baseline values.

View larger version (12K): [in this window] [in a new window]   Figure 8.   Effect of BMS182874 on LPS-induced changes in lung dynamic compliance. Data are expressed as the mean ± SEM. The 2-h point is the minimum value. The 90-min point represents the mean of the data over the second hour. Open circles represent LPS alone experiments. Solid circles are LPS plus BMS182874 experiments. Solid squares are BMS182874 alone experiments. n = 6. An asterisk indicates that the data point is significantly different (p < 0.05) from baseline, and a diamond indicates that the data point is significantly different (p < 0.05) from the time-matched LPS plus BMS182874 data point.

View larger version (13K): [in this window] [in a new window]   Figure 9.   Effect of BMS182874 on LPS-induced changes in AaPO2. Data are expressed as the mean ± SEM. Open circles represent LPS alone experiments. Solid circles are LPS plus BMS182874 experiments. Solid squares are BMS182874 alone experiments. n = 6. An asterisk indicates that the data point is significantly different (p < 0.05) from baseline.

BMS182874 Alone

With the infusion of BMS182874, P_PA declined within 15 min of beginning infusion of the loading dose and then remained slightly but significantly lower than baseline throughout the duration of drug infusion (Figure 1). BMS182874 also caused P_SA to decline from 81 ± 1 to 69 ± 2 mm Hg 1 h into drug infusion and then remained significantly below baseline throughout drug infusion (Figure 2). The sheep showed no obvious untoward effects from BMS182874 infusion. There were no significant changes in Q_L, L/P, C_LP, C_dyn, P_LA, cardiac output, AaPO₂, pH, WBC, or thromboxane B₂ levels as a result of BMS182874. When 30 µg ET-1 was given as a bolus into the pulmonary artery at the end of the 6 h BMS182874 infusion, no change in pulmonary hemodynamics or lung mechanics resulted.

BMS182874 and LPS

As shown in Figure 1, BMS182874 significantly attenuated the early LPS-induced pulmonary hypertension and completely abolished the late pulmonary hypertension. The sustained decrease in P_SA caused by BMS182874 infusion was not altered by LPS infusion (Figure 2).

In the presence of BMS182874, the LPS-induced increase in Q_L was of lesser magnitude and delayed when compared to the LPS alone experiments (Figure 3). Q_L did increase from 2.1 ± 0.3 to a maximum of 4.6 ± 0.8 ml/15 min 4 h after beginning LPS. L/P ratio rose within the first hour after LPS and reached a peak ratio of 0.83 ± 0.06 2 h after LPS as compared to a baseline of 0.64 ± 0.02 (Figure 4). C_LP gradually rose from 1.56 ± 0.25 to 4.16 ± 0.76 ml/15 min 3 h after LPS (Figure 5).

In the presence of BMS182874, plasma thromboxane B₂ levels increased within 1 h after LPS from 262 ± 78 to 2,124 ± 207 pg/ml (Figure 6). The increase in lung lymph thromboxane B₂ closely paralleled that seen in the LPS alone experiments, with levels increasing from 385 ± 84 to 3,748 ± 1,935 pg/ml with rapid return to baseline (Figure 7).

As shown in Figure 8, BMS182874 completely blocked the LPS-induced decline in C_dyn. There was also no significant increase in AaPO₂ or pH throughout the experiment. WBC declined from 9,484 ± 819/mm³ to a nadir of 4,675 ± 944, which was similar to the decline seen with LPS alone, P_LA and cardiac output did not change significantly in the BMS182874 and LPS experiments.

U46619 and BMS182874

Figure 10 shows the dose response curve for U46619 and P_PA with and without BMS182874 treatment (n = 2). Without BMS182874, P_PA increased to a maximum of 53 cm H₂O at a dose of 1.0 µg/kg/min and then gradually declined to 43 with further increases in dosage. In the presence of BMS182874 there was a blunted P_PA response with a nonparallel shift of the dose response curve to the right, with a maximum P_PA of 43 at 10 µg/kg/min.

View larger version (10K): [in this window] [in a new window]   Figure 10.   Dose response curve for U46619 with and without BMS182874. Solid squares represent experiments with U46619 alone, and open circles represent experiments with U46619 infusion in the presence of BMS182874. n = 2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The mechanism by which pulmonary hypertension is related to a poor prognosis in patients with sepsis and ARDS is not known. The predominant pathophysiologic process involved in the development of ARDS is injury to the pulmonary microvasculature with resultant noncardiogenic pulmonary edema. Pulmonary hypertension may exacerbate the injury process and accelerate the formation of pulmonary edema in the setting of a porous injured pulmonary microvasculature.

In the current experiments, BMS182874 significantly reduced both the early and late pulmonary hypertension observed after endotoxemia. Although endothelin levels are increased by LPS both in vitro and in vivo (16, 17, 19), data from this institution indicate that ovine plasma ET-1 levels do not significantly increase until 60 min after LPS administration and peak at 120 min (18). This delay may reflect the fact that endothelin production appears to be transcriptionally regulated (34); there has been no report to date of significant intracellular or extracellular endothelin storage. Given the rapidity of onset of the early pulmonary hypertension after LPS administration (less than 30 min), it seems unlikely that endothelin per se is responsible. The time course of the late phase pulmonary hypertension seems more amenable to influence by endothelin.

A large amount of experimental data indicates that the early pulmonary hypertension in endotoxemia is due either to thromboxane A₂ or to a closely associated product of arachidonic acid metabolism. Levels of thromboxane B₂ (a stable metabolite of thromboxane A₂) in plasma and lung lymph increase acutely after LPS administration and then rapidly return to baseline (6, 8, 38). Pretreatment with a cyclooxygenase inhibitor (7, 8) or a selective thromboxane synthase inhibitor (9) blocks both this acute rise in thromboxane B₂ levels and the early pulmonary hypertension associated with endotoxemia but does not block the late pulmonary hypertension or the late increases in lung fluid and solute exchange. Specific prostaglandin endoperoxide receptor antagonists also block the early pulmonary hypertension (14).

BMS182874 does not inhibit the production of thromboxane B₂ and therefore does not function as a cyclooxygenase or thromboxane synthase inhibitor. BMS182874 does counteract the effects of thromboxane as indicated by the drug shifting the P_PA-U46619 dose-response relationship "rightward" toward higher doses of U46619. It is likely that the ability of BMS182874 to counteract the effects of thromboxane is the explanation for the drug's ability to block the early pulmonary hypertension observed following endotoxemia.

The previously mentioned studies with interventions that inhibit the production or action of thromboxane all indicate that thromboxane is not responsible for the late pulmonary hypertension observed after endotoxemia. None of these interventions alone had any significant effect on the late pulmonary hypertension or the late increases in pulmonary vascular permeability. In the current studies, exogenously administered ETs cause pulmonary hypertension, and an ET receptor antagonist blocks the late LPS-induced pulmonary hypertension. Results reported here suggest that endothelin is responsible for the late pulmonary hypertension, particularly in light of the apparent coincidence of endogenous ET production with the late pulmonary hypertension seen following endotoxemia.

BMS182874 effectively prevented the typical reduction in dynamic compliance of the lungs and increase in AaPO₂ seen with endotoxemia. These results mirror those seen with cyclooxygenase inhibitors and thromboxane receptor antagonists. Although endothelin is an ovine bronchoconstrictor both in the current experiments and those of others (39), it is likely that the effects of BMS182874 on LPS-induced changes in lung mechanics are due to its ability to counteract the effects of thromboxane. Cyclooxygenase inhibitors, thromboxane synthase inhibitors, and thromboxane receptor antagonists have similar effects on LPS-induced changes in lung mechanics and oxygenation. The dramatic changes in dynamic compliance of the lungs after LPS begin more quickly (within 30 min) than transcriptionally regulated endothelin production would likely occur.

Exogenously administered ET-1, ET-2, and ET-3 caused qualitatively and quantitatively similar changes in P_PA and C_dyn. BMS182874 in doses sufficient to block the effects of intravenous boluses of exogenous ET-1 and ET-2 on P_PA and C_dyn also blocked the effects of ET-3. Though BMS182874 appears to selectively inhibit the ET_A receptor in vitro, the drug did not, at the doses used for these experiments, differentially block the effects of ET-1 and ET-2 versus ET-3. We were thus not able to functionally distinguish between ET_A and ET_B receptors in vivo in sheep. These results raise the issue, at least in sheep, that the molecular and in vitro distinction between ET receptor subtypes may not be functionally separable in vivo.

In summary, administration of BMS182874 to sheep lowers systemic and pulmonary arterial pressures and attenuates the pulmonary vascular and lung dynamic compliance changes in response to endotoxemia. These effects most likely result from a combination of non-selective endothelin receptor antagonism and counteraction of the effects of thromboxane. This combination of pharmacologic properties may make this an effective therapeutic agent in ARDS but lessens its usefulness as an experimental probe to define the various roles of endothelin. This notwithstanding, our data strongly suggest that the late pulmonary hypertension associated with endotoxemia is mediated by endothelin.