INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The lung is an important target organ for lipopolysaccharides (LPS) derived from gram-negative bacteria. In recent years it has become clear that overactivation of the immune system by LPS or other inflammatory stimuli may lead to pulmonary complications, and in the worst case to adult respiratory distress syndrome (ARDS). Enhanced pulmonary vascular resistance, as well as bronchoconstriction and airway hyperreactivity (AHR), are important clinical features of ARDS, and contribute to the pulmonary complications of this syndrome (1, 2). Furthermore, AHR is a hallmark of byssinosis (3), and the severity of asthma has been related to the content of LPS in house dust (4). In addition to causing AHR, LPS can also cause pulmonary vascular hyperresponsiveness (VHR) in animal models (5), and might thereby enhance pulmonary hypertension during septic ARDS (1). The mechanisms responsible for both the VHR and the AHR induced by LPS are largely unknown.

Recently, Lefort and colleagues (6) reported that systemic LPS administration induces AHR in C57BL/6 mice. They showed that AHR occurred independently of pulmonary neutrophil recruitment or tumor necrosis factor (TNF) production; however, the mechanism for its occurrence was not resolved (6). Since we have recently shown that the thromboxane receptor agonist U46619 in subthreshold concentrations causes bronchopulmonary hyperreactivity and VHR in murine lungs (7), we examined whether activation of the thromboxane/endoperoxide (TP) receptor is involved in LPS-induced pulmonary hyperresponsiveness.

Here we report that LPS-elicited AHR to methacholine (MCh), as well as pulmonary VHR to platelet-activating factor (PAF), is mediated by the TP receptor. However, the cyclooxygenase (COX) pathway does not exclusively account for generation of the responsible TP agonist(s) in this model. In addition to COX-dependent prostanoid generation, lipid peroxidation by reactive oxygen species (ROS) may participate in the generation of isoprostanes, which are capable of inducing AHR via the TP receptor.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Mice

BALB/C mice of either sex were obtained from Charles River Laboratories (Sulzfeld, Germany) or from the breeding house of the Research Center Borstel. All animals were used at a weight of 18 to 25 g.

Materials

Pentobarbital sodium (Nembutal, Narcoren) was purchased from the Wirtschafts-genossenschaft Deutscher Tierärzte (Hannover, Germany); low-endotoxin-grade bovine serum albumin (BSA) was purchased from Serva (Heidelberg, Germany); acetylsalicylic acid (ASA), actinomycin D (AcD), cycloheximide (CHX), indomethacin, LPS (Salmonella enterica serovar Minnesota, S-form), MCh, N-acetylcysteine (NAC), and PAF (1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine) were purchased from Sigma Chemical (Deisenhofen, Germany); SQ29.548 (1S- [1,2(Z),3,4] -7-[3- [{2- [(phenylamino) -carbonyl] hydrazino} methyl]-7-oxabicyclo[2.2.1]hept-2-yl]-5-heptenoic acid]) NS-398 (N-[2-cyclohexyloxy-4-nitrophenyl] methanesulfonamide), and 8-epiprostaglandin F₂ (8-epi-PGF₂) were purchased from Cayman Chemicals (Ann Arbor, MI); and RPMI-1640 was purchased from BioWhittaker (Verviers, Belgium). LPS from S. enterica serovar Friedenau (S-form) was a kind gift of Helmut Brade (Research Center Borstel).

Isolated Perfused Mouse Lung Preparation

The mouse lungs were prepared and perfused essentially as recently described (8). Briefly, lungs were perfused in a nonrecirculating fashion through the pulmonary artery at a constant flow of 1 ml/min, resulting in a pulmonary artery pressure (Ppa) of 2 to 3 cm H₂O. As a perfusion medium we used RPMI-1640 lacking phenol red (37° C) and containing 4% low-endotoxin-grade BSA. The lungs were ventilated at negative pressure (3 to 9.5 cm H₂O) at a rate of 90 breaths/min, resulting in a tidal volume of about 220 µl. Hyperinflation (25 cm H₂O) was performed at 5-min intervals. Artificial thorax chamber pressure was measured with a differential pressure transducer (DP 45-24; Validyne, Northridge, CA), and airflow velocity was measured with a pneumotachograph tube connected to a differential pressure transducer (DP 45-15; Validyne, Irvine, CA). The arterial pressure was continuously monitored by means of a pressure transducer (Isotec Healthdyne) that was connected with the cannula ending in the pulmonary artery. All data were transmitted to a computer and analyzed with Pulmodyn software (Hugo Sachs Elektronik, March Hugstetten, Germany). The data for lung mechanics were analyzed by applying the following formula:

where P is chamber pressure, C is pulmonary compliance, V is volume, and R_L is pulmonary resistance. The measured R_L was corrected for the pneumotachometer and tracheal cannula resistance of 0.6 cm H₂O · s/ml.

Experimental Design

After preparation, the lungs were perfused and ventilated for 45 min in order to obtain a baseline state. In the experiments addressing LPS-elicited AHR, LPS (5 ng/ml to 50 µg/ml) was added after this control period and was perfused through the lungs for another 210 min. MCh (10 µM) was perfused at 20 min before the addition of LPS, in order to obtain a reference value for bronchoconstriction, and also at different time points after the addition of LPS for periods of 10 min each to determine AHR. In a separate set of experiments, PAF-induced vasoconstriction was investigated. In these experiments, lungs were perfused with 50 µg/ml LPS for 60 min before PAF (250 nM) was added to the buffer, and the subsequent vasoconstriction was followed for 20 min. AcD (300 nM), ASA (500 µM), CHX (100 µM), and NAC (5 mM) were added 45 min before, and indomethacin, SQ29.548, and NS-398 (10 µM, each) were added 10 min before the addition of LPS to the perfusion buffer. 8-epi-PGF₂-elicited AHR was assessed 10 min after the administration of 8-epi-PGF₂. During the experiments, perfusate samples were taken every 10 min and immediately frozen. At the end of the experiments, some of the lungs were flash frozen in liquid nitrogen for subsequent isolation of RNA. Lung tissue and perfusate samples were stored at 80° C.

Reverse Transcription-Polymerase Chain Reaction

Total RNA from lung tissue specimens of 50 to 80 mg was isolated by using Trizol reagent (Gibco, Eggenstein, Germany) according to the supplier's instructions. Four micrograms of total RNA were used for reverse transcription (RT) with Superscript II (Gibco) and oligodeoxythymidine (oligo-dT)-primers. Polymerase chain reaction (PCR) amplification was done with the complementary DNA (cDNA) templates of the mRNAs of interest with the following primer pairs: COX-2: GATCATAAGCGAGGACCTGG (start: 695 bp) and CTGCTTGTACAGCAATTGGC (start: 1,397 bp); -actin: AGACTTCGAGCAGGAGATGG (start: 743 bp) and CAACGTCACACTTCATGATGG (start: 942 bp). The reactions were cycled 35 times (60 s at 94° C, 60 s at 58° C, and 90 s at 72° C after a 5 min denaturing step at 95° C). Products were analyzed by 2% agarose gel electrophoresis and ethidium bromide staining.

Measurement of Prostanoids

Thromboxane B₂ (TXB₂; the stable metabolite of TXA₂) and 8-epi-PGF₂ were measured in perfusate and bronchoalveolar lavage fluid (BALF) samples with enzyme immunoassays (Cayman Chemicals) according to the supplier's instructions. The detection limits for TXB₂ and 8-epi-PGF₂ were 10 pg/ml and 6 pg/ml, respectively.

Statistics

Data in the figures are given as mean ± SEM and those in the text as mean ± SD. The data in Figures 4 and 5 (antagonists versus VHR) and in Figure 9 (AHR elicited by 8-epi-PGF₂) were evaluated by one-way analysis of variance with Tukey's test, using GraphPad Prism version 3.0 software for Windows 95 (GraphPad Software, San Diego CA; ). The data in Figure 2 (LPS concentration response) and Figures 6 and 7 (antagonists versus AHR) were analyzed with repeated measures ANOVA followed by individual contrasts (JMP 3.2.2; SAS Institute, Cary NC). The  level for the multiple contrasts was adjusted by using Hommel's procedure (9).

View larger version (17K): [in this window] [in a new window]   Figure 4.   Effects of ASA, NS-398, and SQ29.548 on LPS-elicited VHR. Columns represent the time integrals of the increase in Ppa during a 20-min perfusion period with 250 nM PAF (VC = vasoconstriction) in the absence (control; n = 4) or presence (LPS; n = 4) of 50 µg/ml LPS (pretreatment for 60 min) in the perfusion buffer. SQ29.548 (10 µM; n = 3) and NS-398 (10 µM; n = 3) were added 10 min and ASA (500 µM; n = 3) was added to the perfusion buffer 30 min before the administration of LPS. *p < 0.05 versus control; #p < 0.05 versus LPS.

View larger version (20K): [in this window] [in a new window]   Figure 5.   Effects of AcD and CHX on LPS-elicited VHR. Columns represent the time integrals of the increase in Ppa during a 20-min perfusion period with 250 nM PAF (VC = vasoconstriction) in the absence (control; n = 3) or presence (LPS; n = 3) of 50 µg/ml LPS (pretreatment for 60 min) in the perfusion buffer. AcD (300 nM; n = 3) and CHX (100 µM; n = 3) were added 45 min before the administration of LPS to the perfusion buffer. *p < 0.05 versus control; #p < 0.05 versus LPS.

View larger version (20K): [in this window] [in a new window]   Figure 9.   AHR elicited by 8-epi-PGF2 in the perfused mouse lung. Columns represent the time integrals of the increase in RL during a 10-min perfusion period with 10 µM MCh (BC = bronchoconstriction) at 20 min before (control, n = 11) and 10 min after the addition of 100 nM (n = 2), 300 nM (n = 3), or 1 µM (n = 3) 8-epi-PGF2. SQ29.548 (10 µM) was added to the perfusion buffer 10 min before 1 µM 8-epi-PGF2. Data are given as mean ± SEM. *p < 0.05 versus control.

View larger version (28K): [in this window] [in a new window]   Figure 2.   Concentration-dependency of the induction of AHR by LPS. Data are normalized to the reference challenge with MCh (10 µM) (BC = bronchoconstriction) conducted 20 min before the addition of LPS to the perfusion buffer, and are expressed as the time integral of the increase in RL during a 10-min perfusion period with MCh (Figure 1). Data are given as mean ± SEM, n = 3 for all groups except the control (n = 7) and 50 µg/ml LPS (n = 6) groups. *p < 0.05 versus control.

View larger version (26K): [in this window] [in a new window]   Figure 6.   Effects of ASA, indomethacin, NS-398, and SQ29.548 on LPS-elicited AHR. Data are normalized to the reference challenge with MCh (10 µM) conducted 20 min before the addition of LPS to the perfusion buffer, and are expressed as the time integrals of the increase in RL during a 10-min perfusion period with MCh (BC = bronchoconstriction; Figure 1). Lungs were perfused either with 50 µg/ml LPS alone (n = 10) or in combination with ASA (500 µM; n = 4), indomethacin (10 µM; n = 4), NS-398 (10 µM; n = 4), or SQ29.548 (10 µM; n = 4), with indomethacin, NS-398, and SQ29.548 being added to the buffer at 10 min and ASA at 45 min before LPS. Control lungs (n = 7) were perfused without LPS. *p < 0.05 versus control; #p < 0.05 versus LPS.

View larger version (23K): [in this window] [in a new window]   Figure 7.   Effects of AcD, CHX, NAC and a combination of NAC and NS-398 on LPS-elicited elicited AHR. Data are normalized to the reference challenge with MCh (10 µM) conducted 20 min before the addition of LPS to the perfusion buffer and are expressed as the time integrals of the increase in RL during a 10-min perfusion period with MCh (BC = bronchoconstriction; Figure 1). In the different groups, lungs were either perfused with 50 µg/ml LPS alone (n = 10) or in combination with AcD (300 nM; n = 3), CHX (100 µM; n = 5), NAC (5 mM; n = 3), or 5 mM NAC + 10 µM NS-398 (n = 4), with NS-398 being added to the buffer at 10 min and AcD, CHX and NAC at 45 min before LPS. Control lungs (n = 7) were perfused without LPS. *p < 0.05 versus control; #p < 0.05 versus LPS.

View larger version (16K): [in this window] [in a new window]   Figure 1.   Time course of lipopolysaccharide (LPS)-induced airway hyperreactivity (AHR) in perfused mouse lung. Lungs were perfused and ventilated for 45 min before LPS (50 µg/ml) was added to the perfusion buffer (t = 0). Methacholine (MCh) (10 µM) was given 20 min before the administration of LPS to obtain a reference value, and again at 60, 120, and 180 min after the addition of LPS (for 10 min each) to determine AHR. Control lungs received no LPS.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Time Course and Concentration-Response Relationship of Endotoxin-Elicited Airway and Vascular Hyperreactivity

Perfusion of murine lungs with 50 µg/ml LPS, a concentration typically found in the serum of rodents subjected to LPS shock (10), did not increase R_L within the observation period of 210 min (Figure 1). However, the airway response to MCh (10 µM) was potentiated by LPS in a time-dependent manner (Figure 1), resulting in an approximately fourfold increase in MCh-elicited bronchoconstriction after 180 min of LPS perfusion, whereas the bronchoconstriction elicited by MCh in control lungs increased only slightly over the experimental time (Figure 1). In separate experiments, we focused on the early time course of LPS-elicited AHR. AHR was detectable as soon as 30 min after the administration of LPS but was not yet visible 5 min after LPS perfusion (data not shown). The hyperreactivity following LPS showed a clear concentration dependency, with the maximal effect being achieved at a concentration of 50 ng/ml LPS in the perfusion buffer (Figure 2). Most experiments were performed with LPS from S. enterica serovar Minnesota. However, to show that the LPS-induced hyperresponsiveness was not specific for this particular LPS, we also demonstrated similar AHR in lungs treated with LPS from S. enterica serovar Friedenau (data not shown).

In addition to LPS-elicited priming of the airway smooth musculature, we investigated the effects of LPS on the pulmonary vasculature. LPS alone did not induce vasoconstriction in the perfused mouse lung within 210 min (data not shown). Perfusion with PAF (250 nM) caused moderate vasoconstriction, with a maximal increase in Ppa of 6.2 ± 0.8 cm H₂O (n = 4) in control lungs (Figure 3). This PAF-elicited vasoconstriction was strongly enhanced when lungs were pretreated with 50 µg/ml LPS for 60 min, resulting in a maximal rise in Ppa of 19.0 ± 1.5 cm H₂O (n = 4).

View larger version (21K): [in this window] [in a new window]   Figure 3.   LPS-induced vascular hyperreactivity (VHR) in response to platelet-activating factor (PAF). Lungs were perfused for 60 min with LPS (50 µg/ml) before PAF (250 nM) was added to the perfusion buffer (t = 0). Data are expressed as the mean ± SEM increase in Ppa (PAP) elicited by PAF in the presence (n = 4) or absence (n = 4) of LPS in the perfusion buffer.

Role of the COX Pathway and the TP Receptor in LPS-Elicited VHR

To investigate the participation of COX products in LPS-elicited VHR, we treated mouse lungs with the nonspecific COX inhibitor ASA (500 µM), the COX-2 inhibitor NS-398 (10 µM), or the TP-receptor antagonist SQ29.548 (10 µM). LPS-elicited VHR was completely abrogated by blockade of the TP receptor with SQ29.548 (Figure 4). In addition, nonspecific inhibition of cyclooxygenases with ASA or specific inhibition of COX-2 with NS-398 provided complete protection, indicating that LPS-elicited VHR is mediated by a COX-2-dependent agonist of the TP receptor.

In another set of experiments, we investigated the contribution of COX-2 gene expression and protein biosynthesis to LPS-elicited VHR. Pretreatment of mouse lungs with AcD (300 nM) or CHX (100 µM) attenuated VHR at 60 min after LPS challenge (Figure 5). Although inhibition of translation by CHX provided complete protection from LPS-elicited VHR, only partial protection was achieved by inhibiting DNA transcription with AcD.

Role of the COX Pathway and the TP Receptor in LPS-Elicited AHR

Following our investigation of LPS-elicited VHR, we tested the effect of the TP-receptor antagonist SQ29.548 on LPS- induced AHR. SQ29.548 (10 µM) completely abrogated LPS-induced AHR at 60, 120, and 180 min after the addition of endotoxin (Figure 6). We therefore expected similar protection by blocking COX activity as observed with vascular reactivity. However, LPS-elicited AHR was only delayed and not completely blocked by the COX-2 inhibitor NS-398. In addition, the nonspecific COX inhibitors ASA (500 µM) and indomethacin (10 µM) failed to protect against LPS-elicited AHR. In contrast to their vascular effects, inhibition of gene transcription by AcD or of protein biosynthesis by CHX showed little efficacy at 60 and 120 min after LPS challenge; at 180 min after LPS, neither of these two inhibitors protected against LPS-elicited AHR (Figure 7).

Induction of COX-2 mRNA in LPS-Treated Lungs

To investigate the induction of COX-2 by LPS, and the efficacy of AcD in suppressing its induction, we analyzed the expression of COX-2 mRNA with RT-PCR. We observed a basal expression of COX-2 mRNA in control lungs perfused for 270 min (Figure 8). These basal mRNA levels for COX-2 were increased in LPS (50 µg/ml)-treated lungs after 210 min of LPS perfusion. When lungs were treated with 300 nM AcD in addition to LPS perfusion, expression of COX-2 was almost completely inhibited (Figure 8).

View larger version (18K): [in this window] [in a new window]   Figure 8.   RT-PCR analysis of lung tissue obtained after perfusion with LPS for 210 min. Shown are COX-2 and -actin mRNA from lungs perfused for 210 min in the absence (control) or presence (LPS) of 50 µg/ ml LPS (total perfusion time = 270 min). In the AcD/LPS-group, AcD (300 nM) was added to the perfusion buffer 45 min before the addition of LPS.

Effects of the Radical Scavenger NAC on LPS-Elicited AHR

Because endotoxin-elicited AHR, unlike VHR, appeared to be only marginally mediated by a COX-2-derived TP-receptor agonist, we hypothesized that oxygen radicals might lead to the peroxidation of lipids in the lung and thus to a COX-independent formation of isoprostanes, which could then account for the COX-independent part of the LPS-elicited AHR.

We therefore investigated the effects of the radical scavenger NAC on LPS-elicited AHR. When lungs were pretreated with 5 mM NAC, AHR following LPS challenge was partly reduced (Figure 7). Combined treatment of the lungs with 10 µM NS-398 and 5 mM NAC provided nearly complete protection from LPS-elicited AHR only at 60 min after LPS administration. At later time points, the effect on AHR of simultaneous pretreatment with NS398 and NAC was not different from that of treatment with NAC alone.

Induction of AHR by 8-epi-PGF₂

Because the results of pretreatment with NAC suggested the involvement of oxidized fatty acids in LPS-induced AHR, we tested the potency of 8-epi-PGF₂, an isoprostane formed by radical-induced lipid peroxidation, in its capacity to elicit AHR in the perfused mouse lung. Perfusion of lungs with 8-epi-PGF₂ for 10 min elicited AHR to MCh in a concentration-dependent manner (Figure 9). This priming effect on the airways was absent at an 8-epi-PGF₂ concentration of 100 nM, but was present at concentrations of 300 nM and 1 µM. In addition, AHR elicited by 8-epi-PGF₂ was completely abolished by the TP antagonist SQ29.548 (Figure 9).

Additionally, we analyzed the release of prostanoids from LPS-treated lungs. In this procedure, we assayed samples of lung effluent as well as BALF from lungs that were perfused for 210 min with 50 µg/ml LPS for TXB₂ and 8-epi-PGF₂ with an enzyme immunoassay. However, neither TXB₂ nor 8-epi-PGF₂ was detected in the perfusate or in the BALF (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

AHR is a clinical symptom that frequently accompanies a variety of inflammatory lung disorders, such as ARDS (2), byssinosis (3), airway infections (11), and asthma (12). In addition, hyperreactivity of the vascular smooth musculature of the lung could at least partly account for the pulmonary hypertension seen in ARDS (1). In all of these circumstances, evidence has been presented for a contribution of endotoxin (3, 4). In addition, it has been shown that endotoxin itself is sufficient to cause both AHR (6) and VHR (13) in experimental animals, although the exact mechanism for this remained unknown. Here we show that both LPS-induced AHR and VHR depend on activation of the TP receptor. Notably, we observed hyperreactivity at LPS concentrations as low as 50 ng/ml, suggesting that this observation might be of clinical importance. However, although both AHR and VHR originate from activation of the same receptor, the mediators that activate these receptors appear to differ, being COX-2-derived metabolites in the case of VHR and non-COX-derived mediators in the case of AHR. In addition, we report the novel finding that isoprostanes, which represent a class of non-COX-derived TP-receptor agonists, can elicit AHR.

Previously, LPS-induced AHR and VHR were investigated in differing models (i.e., AHR in vivo [6, 14, 15] and VHR in perfused lungs [5]). An enhanced vasoconstrictive reactivity to PAF following LPS challenge has been reported in the isolated perfused rabbit lung (5), in which a correlation between VHR on the one hand and the expression of COX-2 (13) and formation of thromboxane on the other hand was recently described. Our present study shows that similar mechanisms also apply for mice. In addition, however, our experiments with SQ29.548 directly prove the involvement of TP receptors in LPS-induced VHR. Accordingly, the mechanism of LPS- induced VHR in mice and rabbits appears to be very similar to that of LPS-induced bronchoconstriction in rats (13, 16). In all of these settings, LPS induces the expression of COX-2 as the prerequiste for formation of a TP-receptor agonist, most likely thromboxane, but possibly also other COX-derived TP agonists such as PGF₂, PGD₂ or even PGH₂. It should be noted that the inhibition of VHR by AcD or CHX could theoretically also be explained if enzymes up- or downstream of COX were controlled at a transcriptional level. Recently, evidence was presented for a change in the pulmonary expression of thromboxane synthase (17) or in the amount of the protein itself (18) following peritoneal sepsis or endotoxin administration. Therefore, we cannot completely exclude the possibility that enzymes other than COX-2 contribute to LPS-induced VHR.

The consequence of TP-receptor stimulation appears to depend on the concentration of the agonist: at lower concentrations the result is AHR and VHR, whereas at higher concentrations direct contraction of both airway and vascular smooth muscle occurs (7). Thus, the failure to observe direct bronchoconstriction by LPS in mice and rabbits as opposed to rats may be explained by the higher thromboxane production in rats (16).

However, the mechanism of LPS-induced AHR seems to be more complicated than this. Previously, work on LPS-elicited AHR has primarily been done in the guinea pig. In this species, a variety of inflammatory mediators, such as leukotrienes (19, 20), TXA₂ (21), tachykinins (22), or tumor necrosis factor (TNF) (23) have been suggested to contribute to LPS-induced AHR. In the rat, AHR following administration of LPS was ascribed to TNF and invading neutrophils (15). In contrast, Lefort and colleagues reported that LPS-elicited AHR in response to MCh in the mouse in vivo occurred independently of TNF and invading neutrophils (6). This finding is in accord with our results in the blood-free perfused mouse lung, from which blood-derived leukocytes are absent. Therefore, LPS-induced AHR in the perfused mouse lung seems to mimic the situation in vivo, even though several factors such as blood cells, serum proteins (e.g., LPS-binding protein, soluble CD14) or the nervous control of airway caliber might modify or complicate the response to LPS in the whole animal. Although Lefort and colleagues (6) could not explain the mechanism underlying LPS-elicited AHR, we found that in our model AHR was completely blocked by the TP-receptor antagonist SQ29.548. In keeping with this, low concentrations (10 nM) of the TP-agonist U46619 have been found to cause AHR and VHR in the perfused mouse lung (7).

Nevertheless, several findings indicate that the major mediator responsible for LPS-induced AHR is different from TXA₂ or other COX-derived prostanoids. First, no thromboxane was detected in the perfusate or lavage fluid of LPS-treated lungs, although the release of eicosanoids and cytokines (i.e., prostacyclin, TNF, and interleukin-6) from LPS-treated mouse lungs occurs under these conditions (24). Moreover, inhibition of the cyclooxygenases with ASA and indomethacin, specific inhibition of COX-2 with NS-398, inhibition of COX-2 expression with AcD, or inhibition of translation of COX-2 mRNA with CHX, all of which abrogated LPS-elicited VHR, were nearly ineffective in the case of LPS-induced AHR.

This suggests that a TP agonist that is formed independently of COX-1 or COX-2 contributes to LPS-induced AHR. One possible explanation for this phenomenon would be the existence of a third COX isoform that is not inhibited by ASA, indomethacin, or NS-398. However, because inhibition of protein biosynthesis was hardly effective in preventing LPS- induced AHR, at least the induction of such a third COX isoform seems rather unlikely. Another possible explanation is COX-independent generation of the responsible TP agonist(s). In 1990, Morrow and coworkers reported the discovery in vivo in humans and in rats of a series of prostaglandinlike compounds that were produced independently of the COX enzymes by radical-catalyzed peroxidation of arachidonic acid (25). These so-called isoprostanes are formed under conditions of oxidative stress (26), and increased levels of the isoprostane 8-epi-PGF₂ are found in patients with diseases associated with oxidative stress, such as hepatorenal syndrome, acute liver failure associated with acetaminophen overdose, or scleroderma (27). More recently, increased isoprostane-levels have been reported in the breath condensate or BALF of patients with ARDS (28), asthma (29), or various interstitial lung diseases (30). At least two of the possibly formed isoprostanes, 8-epi-PGF₂ and 8-epi-PGE₂, are biologically active. Both compounds are potent constrictors of the renal vasculature (25, 26), and 8-epi-PGF₂ was shown to elicit vaso- and bronchoconstriction in the rat lung (31) by activation of the TP receptor. Because LPS elicits the formation of ROS (30, 33), we hypothesized that peroxidized fatty acids might contribute to LPS-elicited AHR in the mouse lung. Indeed, the radical scavenger NAC provided partial protection from LPS-elicited AHR. Moreover, combined treatment of mouse lungs with the COX-2 inhibitor NS-398 and NAC completely abrogated LPS-elicited AHR at 1 h after LPS administration, even if this synergistic effect did not persist.

In accord with our hypothesis that isoprostanes at least partly mediate LPS-induced AHR, we showed that perfusion of murine lungs with the isoprostane 8-epi-PGF₂ caused AHR in a concentration-dependent manner via activation of the TP receptor. Since TP-receptor activation with U46619 has also been reported to cause AHR in response to MCh (7), the induction of AHR appears to be a consequence of TP- receptor activation that is not agonist dependent.

Our failure to detect 8-epi-PGF₂ in lung effluent or BALF from LPS-perfused lungs suggests that either its concentration was too low or that this particular isoprostane may not be the TP agonist responsible for AHR. Theoretically, 64 different F₂-type isoprostanes can be formed by radical-induced lipid peroxidation (27), of which 45 were detected in the liver of CCl₄-treated rats (36). Moreover, the generation of D₂/E₂-isoprostanes, as well as of isothromboxanes, has been reported in vitro and in vivo (27). Thus, identification of the putatively formed isoprostanes in LPS-treated mouse lungs, and the determination of their potency in eliciting AHR and VHR via activation of the TP receptor, need further investigation.

Taken together, our findings in this study show that LPS causes airway and pulmonary vascular hypereactivity in the perused mouse lung via activation of the TP receptor. However, generation of prostanoids that depend on induction of COX-2 can explain only LPS-induced VHR, and not LPS- induced AHR. For the latter, we provide initial evidence for the involvement of free radical-catalyzed arachidonic acid metabolites, suggesting that isoprostanes may participate in LPS- induced lung injury in the mouse.