Cytochrome P450 Metabolites of Arachidonic Acid but Not Cyclooxygenase-2 Metabolites Contribute to the Pulmonary Vascular Hyporeactivity in Rats with Acute Pseudomonas Pneumonia

Experimental Procedures

All animals used in this study were cared for following the principles and guidelines of the Canadian Council on Animal Care and were supervised by a veterinarian. In addition, the ethics review committee at the University of Western Ontario (London, ON, Canada) approved all protocols.

Acute Pneumonia Model

The acute pneumonia rats were prepared as described previously (Yaghi and McCormack, 1999). Briefly, male Sprague-Dawley rats (275–350 g) were randomized to a pneumonia group or control group. Animals in both groups were anesthetized with halothane and a jugular venous line was placed for fluid administration. Animals in the pneumonia group were injected intratracheally with 0.15 ml of saline containing 3 × 10⁸ colony forming units/ml through a tracheotomy. Within 36 h, this instillation of bacteria produced an acute localized pneumonia in the left lung with the right lung appearing grossly normal. Animals in the control group had a tracheotomy only. Postoperatively the rats were housed separately and allowed free access to standard rat chow and water. Fluid was maintained, for all rats, by a continuous infusion of heparinized saline (1 U/ml) at 2 ml/h. Fentanyl, 1.0 μg/ml, was added to the venous infusion for analgesia. Forty-four hours following surgery, rats were anesthetized with pentobarbital (30 mg/kg intravenously), and the thorax was opened. The heart and lungs were removed en bloc and perfused through the pulmonary artery with modified Krebs' solution (see below). Lungs that were used for vascular reactivity studies were placed in cold Krebs' buffer during the dissection of arterial rings. Lungs used for homogenate and microsome preparation were placed in labeled cryogenic vials, sealed, and frozen immediately in liquid nitrogen. All vials were stored at −80°C until used.

In Vitro Vascular Reactivity

Small intrapulmonary arteries (200–400-μm diameter) were dissected from the left lobe of control or pneumonia rats. Vessels were dissected under the light microscope and were cut into cylindrical segments, 1.0 to 2.0 mm in length. Each vessel segment was suspended on two stainless steel wires in 5-ml organ baths (37°C) as described earlier (Yaghi and McCormack, 1999).

The vascular preparations were bathed in a modified Krebs' bicarbonate buffer containing 118 mM NaCl, 4.72 mM KCl, 2.52 mM CaCl₂, 1.2 mM MgSO₄·7H₂O, 1.2 mM KH₂PO₄, 11.1 mM dextrose, 22.1 mM NaHCO₃ (pH = 7.4). Organ baths were continuously gassed with 95% O₂, 5% CO₂. With each agonist, vessels were allowed to contract until a plateau was obtained before adding the next incremental concentration. After maximal contraction with each agonist was obtained, the bath was washed three to four times with fresh Krebs' solution. All vessels were tested for the presence of a functional endothelium by precontracting with PE (7.3 μM) and assessing relaxation to acetylcholine (10.0 μM) at the beginning of each experiment. All vessels used in this study relaxed >50% to acetylcholine.

Assessment of Vasoactive Effects of EETs and 20-HETE.

To study the direct vasoactive effects of CYP metabolites of AA, in vitro vascular responses were assessed by obtaining cumulative concentration-response curves (1 pM–10 μM) to EETs (5,6-EET, 8,9-EET, 11,12-EET, or 14,15-EET) and 20-HETE in comparison to KCl and PE. Separate PA rings were used for each CYP metabolite. In addition, EETs and 20-HETE were assessed for relaxant effects after precontracting PA rings with PE (7.3 μM); none of the compounds exhibited any relaxant effects even at the highest concentrations studied.

Effect of NS-398.

In vitro PA contractility was assessed using three different contractile agonists: KCl, a voltage-dependent agonist, and two receptor-dependent agonists, PE and PGF_2α. NS-398, a selective COX-2 inhibitor, was used to investigate the role of COX-2 products in the depressed vascular contractility observed in pneumonia. Paired pulmonary arteries (n = 26 arterial rings, <400-μm diameter) from the same position in lung and representing adjacent portions of the same vessel, dissected from the affected lobe of the pneumonia-treated rats (n = 7) or the corresponding lung of control rats (n = 6) were equilibrated with vehicle (ethanol; no more than 0.06% v/v) or NS-398 (10 μM) for 30 min. Cumulative concentration-contraction curves for KCl (2–128 mM), PE (0.01–22 μM), and PGF_2α (0.01–200 μM) were obtained as described earlier.

When studying its effects, NS-398 was continuously present in the Krebs' solution throughout the entire experiment. Neither vehicle (ethanol, 0.06% v/v) nor NS-398 had any effect on resting tension in the vessels studied.

Preparation of Rat Lung Homogenates

Lungs from control rats and left lungs from pneumonia rats were homogenized in homogenization buffer (see below). All tissues and solutions were kept on ice during homogenization and aliquoting of lung samples. Aliquots of lung homogenates were stored separately for NOS and NO_X assays.

Measurement of COX Activity

Control and pneumonia lungs were homogenized in Tris buffer (50 mM, pH 7.4, 4°C) containing phenylmethylsulfonyl fluoride (PMSF, 1 mM) and leupeptin (0.2 mM) in a ratio of 5:1 (v/w) (Vane et al., 1994). Lung homogenates were incubated with AA (60 μM) at 37°C for 30 min. The samples were boiled for 2 to 3 min, and then centrifuged at 10,000g for 30 min. The supernatant fractions were aliquoted into vials and stored frozen (−80°C) until used. All supernatants were tested in duplicate at two different dilutions.

Total PG.

Enzyme immunoassay (EIA) was used to determine the total prostaglandin concentration (Prostaglandin Screen kit; Cayman Chemicals, Ann Arbor, MI) in the supernatant fraction prepared from lung homogenate. The antiserum used in this assay has a 100% cross-reactivity with PGE₁, PGE₂, PGF_{1α ,} and PGF_2α; 51.3% with PGF_3α; 43.6% with 6-keto-PGF_1α; 38.4% with 8-iso PGF_2α; 28.5% with 8-iso PGE₂; 26.6% with PGD₂; 20% with 8-iso-2,3-dinor PGF_1α; 9.5% with PGE₃; 5% with thromboxane B₂(TXB₂); and <0.01% each with PGA₁, PGA₃, PGB₁, 15-keto PGE₂, and PGF-M.

PGE₂ and 6-Keto PGF_1α.

The concentrations of two key vasodilator prostaglandins, PGE₂ and PGI₂, were measured by specific EIA kits for PGE₂ and 6-keto-PGF_1α (the by-product of PGI₂) according to the manufacturer's instructions (Amersham Canada Ltd., Oakville, ON, Canada). In the PGE₂ plate assay, the antiserum used has a 100% cross-reactivity with PGE₂, 25% cross-reactivity with PGE₁, and low cross-reactivity with other metabolic products. In the 6-keto-PGF_1α plate assay, the antiserum used has a 100% cross-reactivity with 6-keto-PGF_1α and low cross-reactivity with other metabolic products. The assay kits were validated with known amounts of PGE₂ and 6-keto-PGF_1α standards with a recovery of 96 and 97%, respectively.

COX activity was measured by the formation of total PG, PGE₂, and 6-keto-PGF_1α. EIA results are expressed as nanograms of product per milligram of lung protein per 30 min. The protein concentration of lung homogenate was determined by a microplate modification of the Bradford method, with bovine serum albumin as the standard and homogenization buffer as the blank (Bradford, 1976).

Incubation of Lung Homogenate with COX-2 Enzyme.

Enzyme activity of exogenously added COX-2 was determined in homogenates from pneumonia lungs (pneumonia + COX-2). An aliquot of pneumonia lung homogenate was incubated with or without COX-2 (20 U) in Tris buffer (50 mM, pH 8.0), with 2 mM phenol, 1 μM hematin, and 1 mM EDTA (37°C). The reaction was initiated by addition of AA (60 μM) and was allowed to proceed for 2 min at which time the reaction was quenched by addition of stannous chloride (100 mg/ml solution in 0.1 M HCl). The reaction was allowed to proceed for an additional 10 min. The reaction was terminated, and the supernatant assayed for total PG by EIA as described earlier.

Preparation and Incubation of Lung Microsomes

Lung microsomes were prepared by differential centrifugation as previously described (Bend et al., 1972). Briefly, frozen (−80°C) lungs from control rats, and pneumonia lobes from pneumonia rats were weighed and homogenized in 5 volumes of ice-cold homogenization buffer (1.15% KCl in 50 mM potassium phosphate buffer, pH 7.4), and centrifuged at 14,500g for 20 min (4°C). The supernatant fractions were then centrifuged at 165,480g for 50 min (4°C). The microsomal pellets were resuspended in 2 ml of homogenization buffer and washed by recentrifugation in an ultracentrifuge at 412,160g for 15 min (4°C). Finally, microsomes were resuspended in 1 ml of buffer, divided into 200-μl aliquots, and frozen immediately until needed for incubation. A 50-μl aliquot of the microsomes was used for protein determination by the method of Lowry et al. (1951) using bovine serum albumin as standard.

To study formation of AA metabolites by CYP, lung microsomal protein (1 mg) was incubated with AA (0.2 μCi of [¹⁴C]AA + 20 μM AA) at 37°C for 5 min in 0.5 M potassium phosphate buffer, pH 7.4, in the presence of 1 mM NADPH. The total reaction volume was 1 ml. The incubation was terminated with 50 μl of 1 N HCl. After incubation the metabolites were extracted with ether containing 0.01% butylated hydroxytoluene at acidic pH. The ether was evaporated under a gentle stream of nitrogen and the residue reconstituted in acetonitrile. The metabolites were separated by reverse phase HPLC on a Resolve C18 column using a linear gradient from acetonitrile/water (38:62, v/v) to acetonitrile (100) over 60 min, at a flow rate of 1 ml/min. Radioactivity was monitored using a flow-through detector (Beckman, Fullerton, CA). AA metabolite classes were identified based on their retention times relative to those of standards, and quantitated radiochemically as described earlier (Knickle and Bend, 1994).

Assay of Microsomal HO Activity

Lung microsomes from control and pneumonia rats were prepared as described earlier. The rate of microsomal HO activity was determined essentially as described previously (Sinal et al., 1995) except for the utilization of a recombinant biliverdin reductase (GST-BVR) prepared in this laboratory (Sinal and Bend, 1997). The assay mixture contained 7.5 μg of purified GST-BVR, 1 mg of microsomal protein, and 25 μM hemin. Incubations were performed in the dark at 37°C using 0.1 M potassium phosphate buffer (pH 7.4) and were initiated by adding NADPH to a final concentration of 400 μM. The reaction was terminated 25 min later by placing the reaction tubes on ice. The amount of bilirubin formed was quantitated from the absorbance change at 470 nm relative to 530 nm, using an extinction coefficient of 40 mM⁻¹ · cm⁻². Product formation was verified to be linear with time and protein concentration under these reaction conditions. Values are expressed as picomoles of bilirubin formed per minute per milligram of microsomal protein.

Measurement of NO Metabolites in Homogenate

The metabolic end products of NO (NO₂ ⁻ and NO₃ ⁻, collectively NO_X) were determined in supernatant samples using a chemiluminescence detection method as described earlier (Webert et al., 2000). Homogenate samples were centrifuged at 1000g for 10 min at 4°C and the supernatant assayed for NO_X levels. Samples were referenced to a standard curve generated from NO₃ ⁻standards (50 nM–500 μM, r ² > 0.999).

NO Synthase Assay

NOS activity was quantitated as the conversion ofl-[³H]arginine tol-[³H]citrulline as previously described (Scott and McCormack, 1999). Samples were assayed under three different conditions: 1) calcium/calmodulin, 2) EDTA/EGTA, and 3) EDTA/EGTA + l-NAME (1 mM). Calcium-dependent (constitutive) NOS (cNOS) activity was calculated as the difference between the calcium/calmodulin sample (cNOS + iNOS activities; 1) and the EDTA/EGTA sample (iNOS activity only; 3). Nonspecific radioactivity and metabolism of l-[³H]arginine were accounted for by incubating homogenization buffer or tissue homogenate with l-NAME (1 mM) in the incubation buffer containing EDTA/EGTA (3). Thus, the calcium-independent (inducible) NOS activity was calculated as the l-NAME-inhibitable portion of the activity in the samples with EDTA and EGTA. Resultant enzyme activities were expressed as picomoles of l-citrulline formed per minute per milligram of protein. The protein concentration of the tissue homogenate and washed samples was determined by the Bradford method, with bovine serum albumin as the standard and homogenization buffer as the blank (Bradford, 1976).

Materials

PE, acetylcholine iodine, NADPH hemisulfate salt, AA, Tris buffer, PMSF, leupeptin hemisulfate salt, stannous chloride, and bovine serum albumin were purchased from Sigma Chemical Co. (St. Louis, MO). KCl (BDH Chemicals, Toronto, ON, Canada) solution (2.0 M) was freshly prepared in distilled water when needed. PGF_2α(dinoprost tromethamine) was obtained from Upjohn (Montreal, QC, Canada). Stock solutions of PE (10.0 mM), PGF_2α(10.0 mM), and acetylcholine (10.0 mM) were prepared in distilled water and frozen (−4°C) in aliquots until needed. NS-398 was purchased from Cayman Chemicals and dissolved in ethanol. Cyclooxygenase-2 (ovine) was also purchased from Cayman Chemicals and stored at −80°C until used. All salts for Krebs' solution were purchased from BDH Chemicals. [¹⁴C]AA (NEC-661, purity >99%) was purchased from Mandel Scientific Company Ltd. (Guelph, ON, Canada). 20-HETE and EETs (5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET) were purchased from BIOMOL Research Laboratories, Inc. (Plymouth Meeting, PA), diluted with ethanol, and then divided into aliquots and stored at −80°C until used for organ bath studies. Potassium phosphate monobasic was purchased from Fisher Scientific (Mississauga, ON, Canada). Cytoscint Environmentally Safe scintillation fluid, 1,4-dithiothreitol, PMSF, and Coomassie brilliant blue G250 were purchased from ICN Biomedical Inc. (Mississauga, ON, Canada).l-[³H]Arginine was purchased from Amersham (Oakville, ON, Canada). All other reagents were purchased from Sigma-Aldrich Canada (Mississauga, ON, Canada). Drug concentrations are expressed throughout as final molar concentrations in the organ bath or incubation tube.

Data Analysis

All figures were plotted using Prism (GraphPad Software, Inc., San Diego, CA) and analyzed using the statistics package for Social Sciences (SPSS-PC, version 5.0; SPSS, Inc., Chicago, IL) and GraphPad Instat (GraphPad Software, Inc.). Each log (concentration)-response curve was plotted using nonlinear regression analysis for dose-response curves and the concentrations of agonist required to produce half-maximal contraction (EC₅₀) and maximal developed wall tension values (T _max) were determined. To study the effect of pneumonia on responses of the PA to different contractile agonists, the curves were compared between pneumonia and control groups using repeated measures analysis of variance (MANOVA). The rest of the data was analyzed using unpairedt test, one- or two-way ANOVA as reported underResults section. Results are expressed as mean ± standard error of the mean of n values, wheren = number of rats. A value of p < 0.05 was considered significant.

Results

Microsomal CYP Metabolites of AA

HPLC chromatograms of the [¹⁴C]AA metabolites demonstrated that two major classes of P450 primary metabolites, EETs and 20-HETE, were formed at significantly lower rates in lung microsomes from pneumonia rats compared with microsomes from control rats (Fig. 1). The rates of formation of 20-HETE and EETs in lung microsomes from control and pneumonia rats are shown in Fig. 2. Specific activities of the CYP-dependent formation of 20-HETE and EETs formed were depressed in pneumonia lungs compared with controls by approximately 4- and 2.5-fold, respectively.

In Vitro Vascular Responses of Small PA Rings to CYP Metabolites

PA rings from pneumonia lungs exhibited depressed contractility to KCl and PE in comparison to PA rings from control lungs (Fig.3, a and b, respectively). In addition, contractility to 20-HETE was significantly attenuated in pneumonia compared with control PA rings (Fig. 3c). Similarly, PA rings from pneumonia lungs exhibited depressed contractility to 8,9-EET, 11,12-EET, and 14,15-EET but not to 5,6-EET (Fig.4). However, the EC₅₀ values for 20-HETE and all EETs were significantly lower than the EC₅₀ values for KCl and PE (Table 1), indicating that all EETs and 20-HETE contracted PA rings from control and pneumonia lungs with greater potency compared with KCl and PE. The EC₅₀ values for 8,9-EET, 11,12-EET, and 14,15-EET (but not 5,6-EET) were higher in PA rings from pneumonia compared with control lungs, whereas the T _max values for the EETs were not altered (Table 1). In comparison, theT _max values for 20-HETE (similar to those for KCl and PE) were significantly lower in PA rings from pneumonia compared with control lungs (Table 1).

Effect of NS-398 on Pulmonary Vascular Reactivity in Pneumonia

Acute Pseudomonas pneumonia in rats resulted in significantly depressed pulmonary vascular contractility to KCl, PE, and PGF_2α (Fig.5). NS-398 (10.0 μM) did not alter the responses of control or pneumonia vessels to any of the contractile agonists studied. The vehicle ethanol did not exceed 0.06% v/v of the organ bath volume and had no effect on baseline tone or PA contractile responses to agonists.

Measurement of COX Activity

Total PG.

Lung homogenates from control (n = 8) and pneumonia (n = 10) rats formed similar levels of total PG (Fig. 6a). Therefore, pneumonia did not result in an increase in formation of total pulmonary prostaglandins, as measured by EIA. When pneumonia lung homogenates were incubated with or without COX-2 (20 U), the formation of total PG was significantly higher in pneumonia lung homogenates incubated with COX-2 (pneumonia + COX-2: 2010.9 ± 245.2 ng/mg of protein,n = 4; pneumonia-COX-2: 95.5 ± 33.5 ng/mg of protein, n = 4; p < 0.01). Thus, COX-2 enzyme activity was preserved in these homogenates, eliminating the possibility of endogenous COX-2 inhibitors in the lungs from pneumonia animals.

PGE₂ and 6-Keto PGF_1α Levels.

Lung homogenates from control (n = 10) and pneumonia (n = 12) rats did not show any significant difference in the formation of PGE₂ (Fig. 6b) or of 6-keto-PGF_1α (Fig. 6c).

Microsomal HO Activity

The specific activity of HO was significantly (about 2-fold) higher in lung microsomes from pneumonia compared with control rat lungs (Fig. 7a).

Lung Homogenate NO_X and NOS Levels

NO_X levels (Fig. 7b) and iNOS activity (Fig.7c) were also significantly higher in pneumonia lungs compared with control lungs, approximately 4- and 8-fold, respectively.

Discussion

We have previously described the depressed pulmonary vascular contractility to a number of agonists in rats with acutePseudomonas pneumonia (Yaghi and McCormack, 1999). In this study, we rule out the involvement of COX-2 metabolites and suggest a possible role for CYP metabolites of AA in this phenomenon. Recently, molecular and immunological evidence has shown that isozymes of the CYP2J subfamily are highly expressed in both human and rat lung with prominent expression in vascular smooth muscle cells and endothelium (Zeldin et al., 1996). The CYP proteins and monooxygenase activity required to synthesize 20-HETE are also present in human lung (Birks et al., 1997). Our study confirms the formation of CYP metabolites of AA in rat lung tissues and, for the first time, demonstrates depressed production of EETs and 20-HETE in Pseudomonas pneumonia.

The EETs and 20-HETE have varied vasoactive effects dependent on the vascular beds studied (Leffler and Fedinec, 1997; Gebremedhin et al., 1998). Furthermore, the vasoactive effects of EETs and 20-HETE can be influenced by the activity of other enzyme systems such as NOS and COX (Oyekan and McGiff, 1998; Pratt et al., 1998). Indeed, 20-HETE can act as a cyclooxygenase-dependent vasodilator of human pulmonary arteries (Birks et al., 1997), but acts as a constrictor of rat renal (Sun et al., 1998) and cat cerebral arteries (Gebremedhin et al., 1998). The vascular responses to EETs have been mainly documented as relaxant. EETs act as dilators of renal and cerebral arteries, and the dilator effect is dependent on synthesis of dilator prostanoid analogs (Carroll and McGiff, 1997; Leffler and Fedinec, 1997). In the present study, we demonstrated 20-HETE and EETs to be more potent vasoconstrictors than PE and KCl on small pulmonary arteries. Furthermore, similar to PE and KCl, contractility to these CYP metabolites (except 5,6-EET) was depressed in pulmonary arteries from pneumonia compared with control rats. Our study is the first demonstration of the direct effects of EETs on small pulmonary arteries from normal and pneumonia lungs of the rat. In contrast to what has been previously reported in other vascular beds, all EETs acted as potent constrictors of small pulmonary arteries in concert with the recent findings in rabbit pulmonary arteries by Zhu et al. (2000). The effect of pneumonia on the PA contractility to EETs was different from that to PE, KCl, and 20-HETE in that only the EC₅₀ values of EETs (except 5,6-EET) were higher in pneumonia compared with control arteries, whereas controlT _max values were reached in pneumonia arteries at higher concentrations of EETs. Although no known receptors have yet been identified, both EETs and 20-HETE could exert their contractile effects by modulating the function of potassium and calcium channels in vascular smooth muscle cells (Sun et al., 1998; Mombouli et al., 1999). There was no significant difference in the contractile effect of 5,6-EET on PA rings from pneumonia compared with control lungs, indicating a distinct mechanism of action compared with the other EETs. This could also be due to the fact that in oxygenated Krebs' buffer, 5,6-EET degrades to its 5,6-δ-lactone derivative and 5,6-dihydroxyeicosatrienoic acid (5,6-DHT) (Fulton et al., 1998), thus masking the direct effects of this EET on PA rings.

Nonetheless, all CYP metabolites of AA (20-HETE, 8,9-EET, 11,12-EET, and 14,15-EET) except 5,6-EET were found to be potent pulmonary vasoconstrictors whose contractile effects are depressed in acutePseudomonas pneumonia, indicating that these metabolites may act as modulators of pulmonary vascular tone both in health and disease.

In a previous study on pulmonary arteries isolated from the same rat model of pneumonia, we demonstrated that meclofenamate (1.0 μM), a nonselective inhibitor of COX-1 and COX-2 (Smith et al., 1994), could restore to normal the attenuated contractile response to KCl and PGF_2α (Yaghi and McCormack, 1999). In addition to its role in inflammation (Vane et al., 1994), it has also been suggested that COX-2 may play a physiological role in the regulation of vascular tone in rat lungs (Ermert et al., 1998). In the current study, we used NS-398, a selective COX-2 inhibitor, at a concentration (10 μM) documented to selectively block COX-2 in vitro (Futaki et al., 1994; Huff et al., 1995). Under these conditions, NS-398 did not reverse the attenuated contractile responses to KCl, PE, and PGF_2α in control arteries or in arteries from pneumonia lungs. These data suggest that relaxant PGs produced by COX-2 neither modulate the contractility of control arteries nor contribute to the depressed contractility in arteries from pneumonic lungs. There are a couple of explanations why meclofenamate but not NS-398 reversed the observed depressed contractility. First, our previous observation with meclofenamate could have resulted from a shift in the metabolism of AA toward alternative pathways, including CYP and/or lipoxygenase pathways (Smith et al., 1991). In this case, inhibition of COX enzymes could produce a shift toward increased formation of vasoconstrictor metabolites of these pathways, resulting in a restoration of contraction. Second, an interesting possibility is that meclofenamate indirectly modifies the action of other AA metabolites. Indeed, some CYP metabolites such as 20-HETE and 5,6-EET can be metabolized by cyclooxygenase to vasoactive metabolites, which have been implicated to function as intracellular signaling molecules in arteries (McGiff, 1991; Harder et al., 1997). Thus, inhibition of COX could inhibit further metabolism of 20-HETE and EETs, resulting in the accumulation of these pulmonary vasoconstrictive metabolites and restoration of contractility. This explanation is intriguing since it excludes a role for prostanoid COX metabolites but not COX enzymes in the depressed pulmonary vascular contractility observed in pneumonia.

Incubation of control or pneumonia lung homogenates with AA and measurement of PGs formed by EIA strongly support the results of the in vitro organ bath studies with NS-398. The rate of formation of total PGs, and specifically the relaxant prostaglandins PGE₂ and PGI₂ did not increase in pneumonia lungs compared with control lungs. COX-1 and COX-2 are constitutively expressed in rat lung (Ermert et al., 1998). Therefore, the lack of significant difference in prostaglandin levels formed between pneumonia and control lungs is likely a result of no change in either COX-1 or COX-2 activity in the pneumonia lungs. Another explanation for these results could be that COX-1 is down-regulated, but COX-2 is up-regulated, resulting in a zero net change in prostaglandin levels formed. Indeed, endotoxin treatment in vivo down-regulates COX-1 mRNA expression, but up-regulates COX-2 mRNA expression in rat lung and heart (Liu et al., 1996). In either scenario COX-2 induction does not account for the changes in pulmonary vascular reactivity in this model because NS-398 (a selective COX-2 inhibitor) did not affect baseline tone nor reverse the impaired contractility of PA rings from pneumonia lungs.

The finding that the formation of relaxant prostanoids is not increased in this pneumonia model contrasts to previous work by others (Light, 1986; Hanly et al., 1988). Nevertheless, this difference could be due to differences in species (rat versus human and dog lungs), measurement of circulating plasma PG levels instead of local tissue levels, or the methods of measurement of prostanoids and standardization of tissue PG levels. In our study we used EIA to measure formation of pulmonary tissue prostanoids, something not done in previous studies.

Some studies, using models of infection and inflammation, have demonstrated a role for NO in modulating other enzyme systems such as COX (Vane et al., 1994; Swierkosz et al., 1995). However, using more selective inhibitors of iNOS and COX-2, recent investigations have demonstrated that the two pathways are independently active (Hamilton and Warner, 1998). Indeed, the level of interaction between the two systems is dependent on the model used and on the stage of infection or inflammation (Vane et al., 1994; Salvemini et al., 1996). In the current study, we demonstrated that COX-2 added exogenously is functional in lung homogenates from pneumonia animals. Other studies on acute inflammatory models have also implicated NO as a mediator of decreased CYP monooxygenase activity (Khatsenko and Kikkawa, 1997). NO inhibits CYP activity due to binding of NO to ferrous or ferric iron of the catalytic heme moiety (Wink et al., 1993). NO can also modulate CYP activity by diminishing the mRNA expression of selected isozymes (Wink et al., 1993; Khatsenko and Kikkawa, 1997), and NOS inhibitors are capable of reversing the decreases in activity, protein, and mRNA of some CYP isozymes in rat liver (Khatsenko and Kikkawa, 1997). In addition, NO binding to CYP could release heme and activate HO, resulting in the production of CO, which can inhibit CYP activity (Kim et al., 1995; Maines, 1997). Our rat pneumonia model is an acute localized infection, characterized by increased production of NO by iNOS and elevated HO activity. Therefore, we speculate that NO and/or CO may contribute to the decreased CYP-dependent metabolism and/or depressed EETs and 20-HETE vasoconstrictor effects observed in this model of acute pneumonia.

In conclusion, we have demonstrated two different mechanisms by which CYP metabolites might contribute to the depressed contractility observed in PA rings from rats with acute Pseudomonaspneumonia. The first is the decline in CYP metabolites (EETs and 20-HETE) formed in pneumonia compared with control lungs, and the second is a decline in the direct contractile responses to EETs and 20-HETE in PA rings from pneumonia compared with control lungs. Elevated HO activity, increased NO_X levels, and increased iNOS activity in pneumonia compared with control lungs indicate that NO and/or CO could contribute to the observed CYP metabolic (decreased rate of production of metabolites) and functional (attenuated physiological activity) changes in acutePseudomonas pneumonia. Furthermore, the lack of effect of NS-398 on contractility of PA rings from pneumonia rat lungs, added to the observation that COX activity is not elevated in pneumonia compared with control lungs, indicates that vasodilator prostaglandins produced by cyclooxygenase enzymes, specifically COX-2, do not contribute to the depressed PA contractility in this model of acute pneumonia.