Introduction

Abstract Introduction References

When inflammation induced by microbial infection is excessive and spreads throughout the circulation, a clinical catastrophe can occur. Acute disseminated intravascular inflammation and concomitant failure of multiple organs (1, 2) typify the syndrome. This phenomenon occurs in animals and humans, and is classified as sepsis, severe sepsis, or septic shock. Severe sepsis and septic shock have high mortality rates and are the major causes of death in critical care units in the United States (3). Infections with gram-negative bacteria are common causes of sepsis, severe sepsis, or septic shock, although many other microbes can induce a similar syndrome. Severe sepsis and septic shock are often accompanied by acute inflammatory lung injury, often referred to as the acute respiratory distress syndrome (ARDS) (4). Bacterial lipopolysaccharide (LPS) plays a major role in inducing the syndromes of sepsis, severe sepsis, and septic shock when they occur during the course of infection caused by gram-negative bacteria.

The pathogenesis of these severe inflammatory syndromes partially depends on the induction of inflammatory genes by host leukocytes (1). Two of the essential genes are those for tumor necrosis factor- and interleukin (IL)-1. Other genes that may play crucial roles in severe sepsis are the prostaglandin G/H synthase (PGHS) genes.

Two PGHS genes have been identified (5). Although PGHS-1 is usually constitutively expressed in a number of tissues, including blood monocytes and alveolar macrophages (AM), PGHS-2 is an immediate-response gene that appears to be the more important of the two PGHS genes in regulating acute inflammation (6). The PGHS genes encode proteins that are key rate-limiting enzymes for the metabolism of free arachidonic acid to prostanoids, thromboxane (TX), and prostacylin (5). The two PGHS genes clearly differ. LPS and IL-1 (7) induce the expression of PGHS-2. Adrenocorticosteroids, which are potent antiinflammatory agents that prevent the lethal effects of LPS, inhibit PGHS-2 but not PGHS-1 gene expression (10). PGHS-1 and PGHS-2 can be differently inhibited by pharmacologic agents (11). Recent data suggest that PGHS-1 and PGHS-2 utilize different pools of arachidonic acid (12). Thus, the two PGHS isoforms may play distinct roles in cell function and inflammation. LPS rapidly induces expression of the PGHS-2 gene but not of the PGHS-1 gene in human blood monocytes (13), human AM (14), and rabbit AM (7).

We previously reported that priming of isolated perfused rabbit lung (IPL) for 2 h with LPS, followed by stimulation with platelet-activating factor (PAF), induced an acute TX-dependent lung injury (15). The lung injury was characterized by pulmonary hypertension and lung edema. We later showed that LPS induces expression of the PGHS-2 gene in rabbit AM cultured in vitro. In the study reported here, we tested the hypothesis that expression of the PGHS-2 gene is responsible for the TX-dependent pulmonary hypertension in the LPS-perfused IPL. We conclude that induction of the PGHS-2 gene, and not that for PGHS-1, is responsible for the TX-dependent pulmonary hypertension that occurs when the LPS-primed IPL is stimulated by PAF.

	Materials and Methods

The IPL

The Animal Care and Use Committee of the Wake Forest University School of Medicine approved the protocol for the study. Female New Zealand white rabbits weighing  3 kg were anesthetized with a combination of ketamine hydrochloride (35 mg/kg) and xyline (5 mg/kg) given intramuscularly; additional ketamine was administered intravenously as required. A tracheostomy tube was placed and the lungs were ventilated with a small-animal respirator for the remainder of the experiment. After exsanguinating each rabbit by incising the carotid arteries, the heart and lungs were exposed with a midline sternotomy. The pulmonary circulation was cannulated and perfused with catheters inserted into the pulmonary artery via the right ventricle and into the left atrium through the left ventricle. The pulmonary circulation was perfused with normal saline (NS) containing 25 mg/liter of gentamicin. Initially, 700 ml of NS was perfused through the system and discarded. After this 700-ml washout, a recirculating system was established with 300 ml of perfusate for the duration of the experiment. The initial 700-ml washout effectively removed the animal's plasma and most of the blood cells from the animal's circulation. The recirculating perfusate contained 50 to 250 red blood cells/mm³, 1-2 white blood cells/mm³, and no detectable platelets.

Determining Effects of Actinomycin D and Cycloheximide on LPS-Primed IPL

IPL were perfused for 2 h with NS alone; NS and 100 ng/ml of LPS (Escherichia coli 0111:B4 from Difco), 100 ng/ml of LPS plus 10 µM/ml of cycloheximide (CHX; Sigma Chemical Co., St. Louis, MO), or 100 ng/ml LPS and 100 ng/ml of actinomycin D (AD; Sigma). Either 100 ng/ml PAF or 10 µM arachidonic acid was then added to the perfusate of LPS-primed IPL, nonprimed IPL, and LPS-primed IPL treated with AD or CHX. Pulmonary artery pressure (Ppa) and levels of TXB₂ were monitored at multiple time points for 10 min, as described subsequently.

LPS contamination of IPL was closely monitored in samples of lung tissue and in perfusate at the outset and at the end of experiments, using the Limulus amoebocyte lysate assay (Sigma). Only IPL that remained LPS-free (< 10 pg/ml or per gram of tissue) throughout the experiments were used as controls for the LPS-primed IPL. Animals with LPS contamination at the outset were not studied.

Pressure Monitoring

Ppa was monitored throughout the experiments with pressure transducers in contact with the closed perfusion apparatus and connected to a Grass 7 polygraph machine.

TX Measurement

Samples of perfusate that had been frozen at 70°C were thawed and assayed for TXB₂, a stable metabolite of the active TXA₂. Levels were determined by radioimmunoassay, using specific antiserum to TXB₂ obtained as a generous gift from Dr. Larry Levine of Brandeis University. Samples were assayed in duplicate, and values fell within the linear phase of the standard curve.

RNA Extraction and Northern Blot Analysis

For RNA extraction and Northern blot analysis, a small segment of one lung was isolated by ligating the segment's blood vessels immediately after exposing the lungs, and served as a control. The total lung preparations were then removed, washed with 700 ml of NS as described previously, and then perfused for 2 h with 300 ml of NS containing 100 ng/ml of LPS. Studies of the rabbit AM indicated that 2 h was close to the maximum for expression of PGHS-2 in response to LPS stimulation (7). Tissue sections were recovered, snap-frozen in liquid nitrogen, and shattered. Murine complementary DNA (cDNA) probes for PGHS-1 and PGHS-2 (Oxford Biomedical Research, Oxford, MI) were labeled with ³²P by nick translation (New England Nuclear, Boston, MA). Total RNA was isolated through the RNAzol B method (Tel-Test, Inc., Friendswood, TX). Total RNA (10 µg/lane) was fractionated on a 1% agarose, 6.6% formaldehyde gel in 1 × 3-(N-morpholino) propanesulfonic acid (MOPS) buffer (0.02 M MOPS, 5 mM sodium acetate, 1 mM ethylenediamine tetraacetic acid), and capillary-blotted onto nylon membrane filters (Gene Screen Plus; DuPont, NEN, Boston, MA). UV cross-linking was done with a Stratalinker UV cross-linker (Stratagene, San Diego, CA). Filters were prehybridized in 5 to 10 ml QuikHyb hybridization buffer (Stratagene) for 15 min at 68°C, in an Autoblot Micro Hybridization Oven (Bellco Glass Inc., Vineland, NJ). An aliquot containing 10⁶ cpm probe/ ml QuikHyb and 1 mg salmon sperm DNA were added, and hybridization was done for 1 h at 68°C. The filters were washed twice for 15 min at room temperature in 2× standard saline citrate (SSC)/0.1% sodium dodceylsulfate (SDS), and once for 30 min at 60°C in 0.1× SSC/0.1% SDS. The filters were exposed for 6 to 24 h to Kodak XAR film (Kodak, Rochester, NY), using two intensifying screens, at 70°C. When required, the filters were stripped by boiling for 30 min in 0.1× SSC/0.1% SDS, and reprobed under the same conditions.

Western Blot Analysis

As with messenger RNA (mRNA) analyses, a small segment of one lung was isolated to serve as control by ligating its blood vessels immediately after exposing the lungs. The isolated lung preparations were then removed, washed with 700 ml of NS as described previously, and were then perfused for 2 h with 300 ml of NS containing 100 ng/ml of LPS. To optimize detection of protein, IPL were perfused for 4 h. Control and experimental samples were snap-frozen in liquid nitrogen and shattered. Protein was extracted and separated by 7% SDS-polyacrylamide gel electrophoresis (PAGE), and was then blotted to Hybond-ECL nitrocellulose membrane (Amersham, Arlington Heights, IL) and treated with blocking buffer before overnight incubation at 20°C with either a rabbit polyclonal antipeptide antibody to the unique 17-amino-acid sequence of PGHS-2 (kindly provided by Dr. David DeWitt, Michigan State University) or a polyclonal antibody raised in rabbits to murine PGHS-1 (Cayman Chemical Co., Ann Arbor, MI). The PGHS-1 standard was obtained from Cayman Chemical Co. The PGHS-2 protein standard was from the microsomal fraction of Cos-1 cells transfected with recombinant human PGHS-2 cDNA (supplied by Dr. DeWitt). Proteins were visualized with a horseradish peroxidase-conjugated antirabbit antibody (Cappel; Organon Teknika, West Chester, PA), using enhanced chemiluminescence (DuPont).

Data Analysis

For multiple-time-point experiments (pulmonary artery pressure; TXB₂), data were subjected to one-way or two-way analysis of variance, followed by Duncan's new multiple range test to determine differences between means. Values of P < 0.05 were considered statistically significant. For Northern blot analysis of mRNA or Western blotting, a representative sample of at least five experiments having similar results is shown.

	Results

Effect of AD or CHX on LPS-Primed, PAF-Stimulated IPL

We previously showed that IPL primed for 2 h with 100 ng/ml LPS and then stimulated with 100 ng/ml PAF shows marked increases in Ppa, which depend on increases in circulating TXB₂ that is secreted into the intravascular space rapidly after perfusion with PAF (15). To explore the possibility that transcriptional or translational events might be required for enhancing production of TX, IPL were perfused for 2 h with LPS alone or LPS plus either the transcription inhibitor AD or the protein synthesis inhibitor CHX prior to stimulation of the LPS-primed IPL with PAF. The increases in Ppa and in the level of TXB₂ observed in LPS-primed, PAF-treated IPL were reduced significantly by either AD or CHX (Figures 1A and 1B). This suggests that induction of a phospholipase A₂ or PGHS gene and subsequent protein synthesis are required for the lung injury induced by LPS priming of IPL.

View larger version (14K): [in this window] [in a new window]   Figure 1.   Effects of AD or CHX on responses in Ppa and TXB2 production in IPL primed with 100 ng/ml of LPS and stimulated with 100 ng/ml of PAF. Three groups of lungs were studied: IPL perfused for 2 h with LPS alone and then with PAF (dotted lines with squares); IPL perfused for 2 h with LPS plus 100 ng/ml of AD and then with PAF (dotted lines with circles); and IPL perfused for 2 h with LPS and 10 µM CHX, and then with PAF (dotted lines with diamonds). (A) Changes in Ppa. (B) Changes in levels of TXB2 measured in the perfusate. Data are expressed as means ± SEM of five experiments. *Values with P < 0.005 for IPL perfused with LPS versus IPL perfused with LPS plus either AD or CHX.

Effect of AD or CHX on LPS-Primed IPL Perfused with Arachidonic Acid

Both the release of free arachidonic acid by activation of a phospholipase A₂ and the quantity of PGHS are rate-limiting steps in the production of prostanoids and TX (16). PAF treatment of our model of lung injury in LPS-primed IPL presumably mobilizes free arachidonic acid that is then available for metabolism by a PGHS. To bypass phospholipase A₂ and thereby assess putative PGHS enzymatic activity in LPS-primed IPL, we perfused IPL with 10 µM arachidonic acid rather than with PAF. Either unprimed IPL or IPL primed with LPS at 100 ng/ml were then perfused for 10 min with 10 µM free arachidonic acid (Figures 2A and 2B), and Ppa and TXB₂ were measured. Addition of arachidonic acid to LPS-primed IPL was followed by significant increases in both Ppa and circulating TXB₂, suggesting that there was an increase in the level of PGHS enzymatic activity in the rabbit lung. The response to arachidonic acid was slower than that observed following PAF. Arachidonic acid perfused into unprimed IPL caused no increase in either Ppa or perfusate levels of TXB₂. We then tested whether the increase in PGHS activity in LPS-primed and arachidonic acid-treated IPL could be inhibited by AD or CHX. Figures 2A and 2B show that both of these agents significantly reduced the increased pressure and TXB₂ level in LPS-primed IPL perfused with arachidonic acid. This suggests that induction of a PGHS gene occurs during a 2-h perfusion of IPL with LPS.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Effects of AD or CHX on Ppa response and TXB2 production in IPL primed for 2 h with 100 ng/ml LPS and perfused with 10 µM arachidonic acid (AA). Four groups of lungs were studied: IPL perfused for 2 h with buffer alone and then perfused with arachidonic acid (dotted lines with triangles); IPL perfused with LPS for 2 h and then perfused with arachidonic acid (dotted lines with squares); IPL perfused with LPS and 100 ng/ml of AD and then perfused with arachidonic acid (dotted lines with circles); IPL perfused with LPS for 2 h in the presence of 10 µM CHX and then perfused with arachidonic acid (dotted lines with diamonds). (A) Changes in Ppa. (B) Changes in levels of TXB2 measured in the perfusate. Data are expressed as mean ± SEM of three experiments. *Values with P < 0.05 for IPL primed with LPS and stimulated with arachidonic acid versus the other three groups.

Effect of LPS Priming on Expression of PGHS-1 and PGHS-2 mRNA

To determine whether LPS induces either PGHS-1 or PGHS-2 mRNA in IPL, sections of lung were processed for Northern blot analysis of steady-state levels of PGHS-1 or PGHS-2 mRNA following perfusion of IPL for 2 h with 100 ng/ml of LPS. A segment of a lung ligated prior to perfusion served as an internal control. PGHS-1 mRNA, appearing with the expected 2.8-kb size, was present both in the isolated control segment and following LPS perfusion; no appreciable change in the steady-state level of PGHS-1 mRNA was detected after LPS perfusion (Figure 3). In contrast, the quantity of the 4.8-kb PGHS-2 mRNA increased by 2 h after LPS perfusion (Figure 3). A faint band of PGHS-2 mRNA was visible in the control segment. Contralateral lungs that were perfused with buffer in the absence of LPS also did not express PGHS-2 (data not shown). A time course for induction of PGHS-2 mRNA was not determined, in the interest of reducing the number of animals killed.

View larger version (79K): [in this window] [in a new window]   Figure 3.   Effect of LPS on PGHS-1 and PGHS-2 mRNA in rabbit IPL. A small segment of one lung was isolated to serve as a control immediately after the lungs were exposed by ligating the segment's blood vessels. The lungs were then removed, washed relatively free of blood, and perfused for 2 h with buffer containing 100 ng/ml of LPS. Sections from the control segment and perfused lung were obtained at the end of the perfusion, and the mRNA in both was assessed by Northern blot analysis, using nick-translated probes of murine cDNA for PGHS-1 and PGHS-2. Lane 1: control segment. Lane 2: lung perfused with LPS. The lower panel shows the 28S band as a control for RNA loading. A representative experiment among five is shown.

Effect of LPS Priming on Expression of Immunoreactive PGHS-1 and PGHS-2

PGHS-1 immunoreactive protein was prominent in the control sample and the LPS-primed IPL (Figure 4A). No significant change in PGHS-1 immunoreactive protein was detected during LPS-priming of IPL (n = 7). In contrast, the increase in PGHS-2 mRNA that occurred during LPS perfusion for 4 h was accompanied by an increase in PGHS-2 immunoreactive protein (a representative example is shown in Figure 4B). Although PGHS-2 protein increases in LPS-treated rabbit AM after the 4-h timepoint (7), we were unable to maintain IPL preparations for longer than 4 h.

View larger version (36K): [in this window] [in a new window]   Figure 4.   Effect of LPS on PGHS-1 and PGHS-2 immunoreactive protein following LPS priming of IPL. A small segment of one lung was isolated to serve as an LPS-negative control by ligating its vessels immediately after exposing the lungs. The experimental lungs were then removed, washed with buffer alone, and perfused for 4 h with buffer in the presence or absence of 100 ng/ml of LPS. Proteins were extracted from samples, separated by 7% SDS-PAGE, blotted to Hybond membrane, and probed with a rabbit polyclonal antibody raised against murine PGHS-1 or a rabbit antibody against a 17-peptide sequence unique to PGHS-2. Reactions were visualized by using a horseradish peroxidase-conjugated goat antirabbit antibody and enhanced chemiluminescence. (A) PGHS-1. Lane 1 contains the PGHS-1 standard. Lane 2 represents LPS-negative control. Lane 3 represents the lung stimulated with LPS for 4 h. PGHS-1 appears as a 70-kD band. (B) PGHS-2. Lane 1 contains the PGHS-2 standard. Lane 2 represents LPS-negative control. Lane 3 represents the lung stimulated with LPS for 4 h. PGHS-2 appears as a 72-74-kD doublet. A representative experiment among seven is shown.

	Discussion

This study supports the concept that LPS induction of the PGHS-2 gene in rabbit lung is responsible for the TX- dependent pulmonary hypertension that occurs when the lungs are subsequently stimulated intravascularly with the physiologic agonist PAF. Several observations support this conclusion. First, LPS-induced increases in circulating levels of TXB₂ and Ppa that follow stimulation of LPS-primed IPL with PAF can be inhibited by agents that block gene transcription and translation. Second, perfusion of LPS-primed IPL with arachidonic acid causes increases in Ppa and increases in circulating TXB₂, suggesting that LPS induces an increase in rate-limiting PGHS enzymatic activity. Third, either AD or CHX inhibits the increases in Ppa and TXB₂ observed in LPS-primed IPL that are perfused with arachidonic acid, suggesting that a PGHS gene has been induced during LPS perfusion. Fourth, LPS induction of PGHS-2 mRNA can be detected by 2 h, and PGHS-2 immunoreactive protein within 2 to 4 h. Fifth, and in contrast to the case for PGHS-2, PGHS-1 mRNA and immunoreactive protein are constitutively expressed in IPL, and the levels of the PGHS-1 mRNA and PGHS-1 immunoreactive protein do not change during LPS priming. Sixth, the perfusion of free arachidonic acid into IPL not primed with LPS results in no detectable increase in Ppa or release of TXB₂ into the perfusate.

Surprisingly, although PGHS-1 is constitutively expressed in rabbit lung, increases in perfusate levels of TXB₂ and in Ppa occur only in IPL that have been perfused with LPS for 2 h and then stimulated intravascularly with either PAF or arachidonic acid. A possible explanation for this finding is that exogenous arachidonic acid is metabolized by PGHS-2 rather than by PGHS-1, which is available to arachidonic acid released from intracellular stores (12). Also supporting this concept are our previously published findings that PAF treatment of IPL that have not been primed with LPS produces small but significant increases in Ppa and TXB₂ secretion (15). In this case, PAF alone might be mobilizing arachidonic acid from intracellular stores following stimulation of the PAF receptor and activation of phospholipase A₂. Alternatively, PGHS-1 may be expressed in lung cells that do not contain TX synthase, or in cells inaccessible to arachidonic acid.

An emerging concept is that the PGHS-1 and PGHS-2 gene products may regulate different cellular functions (6). PGHS-1, as a constitutively expressed gene, is available to provide a rate-limiting step in the production of bioactive metabolites that regulate the daily functions of cells. In contrast, PGHS-2 is an immediate-response gene that is tightly regulated and expressed only during special situations, such as acute inflammation (6). The ability to inhibit differentially the two PGHS isoforms provides optimism for advances in the therapy of inflammatory diseases (18, 19).

It is likely that PGHS-2 is an important modifier in inflammatory diseases of the lung, and that products of the metabolism of arachidonic acid by PGHS-2 can be proinflammatory. For example, increases in TX levels and pulmonary hypertension are hallmarks of the ARDS that often accompanies severe sepsis (3, 4, 20). It is possible that the PGHS-2 gene is induced in specific human lung cells when bacterial products such as LPS circulate during severe sepsis and septic shock, and that a PGHS-2-dependent increase in circulating TX plays a pathogenic role in the pulmonary hypertension and lung injury associated with ARDS.

Expression of PGHS-2 also may be associated with antiinflammatory effects. A recent study links repressed expression of PGHS-2 in interstitial lung fibroblasts with the degree of pulmonary fibrosis measured in patients with idiopathic interstitial pulmonary fibrosis (21). Inhibition of PGHS-2 leads to an increase in inflammation in a rat model of colitis (22). AM obtained from rabbits and stimulated with LPS and PAF secrete products that could be both proinflammatory and antiinflammatory (7). It is of interest that in addition to TX, the pulmonary perfusate of rabbit IPL that has been primed and stimulated with PAF contains increased levels of prostaglandin E₂ and D₂ and prostacyclin (unpublished data). However, the biologic vasoconstrictive effects of TX predominate. Thus, the physiologic effects of the products released after the metabolism of arachidonic acid by PGHS-2 may be tissue- or cell-specific.

Our study did not identify the lung cell type(s) responsible for the increase in production of TX. Such cells must both express TX synthase and be responsive to LPS induction of the PGHS-2 gene. Candidate lung cells for these properties include lung tissue cells or AM (7), arterial smooth-muscle cells (23), endothelial cells (7), fibroblasts (24), or mast cells (20). Our attempts to identify the cells with immunohistochemistry were unsuccessful because of the level of nonspecific binding of the anti-PGHS-2 antibodies used in our study.

In summary, circulating LPS induces expression of PGHS-2 in rabbit lung. This enables the lung to respond to other inflammatory stimuli that can mobilize intracellular stores of arachidonic acid by activating one or more of the phospholipase A₂ enzymes. One example of such a putative physiologic or pathophysiologic stimulus is PAF. The presence of an increased level of PGHS-2 and mobilization of arachidonic acid by the second stimulus could then lead to increased production of TX. Increases in levels of circulating TX may cause lung injury by inducing pulmonary hypertension. In contrast to PGHS-2, PGHS-1 is constitutively expressed in rabbit lung but appears not to contribute to the pulmonary hypertensive response that follows LPS perfusion of lung by the intravascular route. Selective suppression of PGHS-2 might have salutary effects in treating human inflammatory diseases such as ARDS.