INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The lung vasculature is known to be a site of intense prostanoid, leukotriene (LT), and hydroxyeicosatetraenoic acid (HETE) biosynthesis (1). LTs are implicated in vasoregulatory and inflammatory events, such as vasoconstrictor responses to various agents and induction of vascular leakage (peptido-LTs), as well as leukocyte chemotaxis and activation (leukotriene B₄; LTB₄). Recently, bioconversion of arachidonic acid (AA) via cytochrome P450 monooxygenases to cis-epoxyeicosatrienoic acids (5,6-, 8,9-, 11,12- and 14,15-cis-EETs) was demonstrated in rabbit, canine, guinea pig, and human lung tissue or microsomal fractions (6). Possible biologic features attributed to these oxiranes include regiospecific vasoregulatory properties (10) and regulation of electrolyte transport processes (2, 10, 12, 13). Interestingly, recent evidence favors the hypothesis that cis-EETs might be intimately related to the so-called "endothelium-derived hyperpolarizing factor" (14).

In buffer-perfused lungs of different species, LTs are known to be liberated into the intravascular space in response to inflammatory stimuli (17). We adapted this technique to perfused human lungs, isolated during surgery for lung neoplasia, and used a newly developed combined technique of multistep, solid-phase extraction (SPE), isocratic reversed-phase high-performance liquid chromatography (RP-HPLC), with online photodiode array detection of eluting compounds, for the simultaneous assessment of all LTs, HETEs, and EETs (20). Calcium ionophore challenge of the human lungs provoked the release of substantial quantities of peptido-LTs and LTB₄, as well as HETEs, into the blood-free perfusate. Most impressively, additional intravascular liberation of all four regioisomeric cis-EETs was noted, with 8,9-EET representing the predominant compound, and the total quantity of EETs surpassed the sum of all LTs and HETEs by more than twofold. We conclude that AA-derived lipoxygenase (LO), and in particular P450 monooxygenase products, are endogenous constituents of the human lung vasculature, suggesting that they have an important role in pulmonary vascular homeostasis and pathophysiology.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Standards, Chemicals, Reagents, Solvents, and Lung Perfusate Media

20-COOH-LTB₄, 20-HO-LTB₄, LTC₄, prostaglandin-B₁ (PGB₁), LTE₄, 6t-LTB₄, 6t,12e-LTB₄, LTB₄, LTD₄, 5S,6R-DiHETE, 5S,6S-DiHETE, 15-HETE, 11-HETE, 8-HETE, 12-HETE, 9-HETE, and 5-HETE were purchased from Biomol GmbH (Hamburg, Germany). LTB₄ -lactone, 5,6-DHET -lactone, 14,15-EET, 11,12-EET, 8,9-EET, and 5,6-EET were supplied by Cascade Biochem Ltd. (Reading, England). All eicosanoids were checked for purity and quantified spectrophotometrically prior to use. The Ca ionophore A23187, butylated hydroxytoluene (BHT), pentanesulfonic acid (PSA), and triethylamine (TEA) were obtained from Sigma Chemie GmbH (Munich, Germany). Ethylenediaminetetraacetic acid (EDTA) and formic acid (FA) were purchased from E. Merck (Darmstadt, Germany). Water was purified with a Milli-Q system (Millipore, Eschborn, Germany). Methanol (MeOH) was delivered by Burdick & Jackson (Muskegon, MI), acetonitrile (MeCN) was purchased from J. T. Baker (Deventer, The Netherlands), and isopropyl alcohol (IPA) was obtained from Fluka AG (Buchs, Switzerland). All solvents used were HPLC-grade or better. The prepared solutions were passed through a 0.2-µm filter (Millipore) and degassed under vacuum. Before use, the solvent mixtures and the mobile phase were additionally degassed by sonication. Krebs-Henseleit buffer was delivered by Serag Wiessner (Naila, Germany).

SPE and HPLC Equipment and Stationary Phases

SPE was performed with ODS cartridges (Chromabond, C18ec) mounted in a Chromabond vacuum manifold, all of which were obtained from Macherey-Nagel (Düren, Germany). Ten-milliliter samples of lung perfusate were extracted on 200 mg/3 ml cartridges (sorbent mass: 200 mg; reservoir volume: 3 ml).

The HPLC equipment consisted of a Gynkotek Model GINA 160 autosampler (Gynkotek GmbH, Munich, Germany) equipped with a 500-µl sample loop and a 250-µl injection syringe; a Gynkotek Model 480 HPLC pump; a Gynkotek Model 320S photodiode array detector; a Waters Model 990 photodiode array detector (Waters/Millipore, Eschborn, Germany), a Shimadzu Model SPD-6A variable wavelength UV detector (Shimadzu, Kyoto, Japan), and two in-line filters (0.2-µm stainless steel frit) from Latek KG (Eppelheim, Germany). The Gynkosoft (Gynkotek) version 5.32E and the Waters 990+ version 5.00 data systems were used to collect the detector output data, as well as for integration, spectral analysis, and processing. All separations were done on an analytical HPLC column (250 mm × 4 mm) filled with ODS-Hypersil (particle size: 3 µm; pore size, 100 Å) obtained from Shandon (Astmoor, UK).

Standard and Quality Control Sample Preparation

In parallel with processing of genuine samples, we analyzed calibration mixtures of authentic standards (Figure 1) and used them for absolute quantitation of the actual set of samples. Moreover, quality control samples containing known amounts of standards dissolved in eicosanoid-free perfusate were processed with each sample batch under the same conditions, allowing determination of the actual recovery for each compound.

View larger version (35K): [in this window] [in a new window]   Figure 1.   RP-HPLC-chromatogram (top panel ) of a 22-component standard mixture with associated spectra plots (bottom panels) provided by an online photodiode array detector.

All calibration standard solutions were prepared from stock solutions in methanol or ethanol by serial dilution with the HPLC mobile phase. Quality control samples were prepared from separate stock solutions of authentic standards and diluted to the final sample volume with the corresponding eicosanoid-free biologic fluid. Calibration standard solutions were stored at 20° C until analysis. Quality control samples were freshly prepared prior to SPE of each sample batch.

SPE and HPLC Procedures

AA metabolites released in human lungs were analyzed as previously described (20), using a method combining solid-phase extraction (SPE), isocratic RP-HPLC separation, and online photodiode array detection (PDAD) and spectrum analysis for identification and measurement of all LTs, HETEs, and EETs within a single run.

SPE. After collection, all samples were supplemented with constant amounts of PGB₁ as an internal standard (IS), and were then centrifuged for 10 min at 1,500 × g and subjected immediately to SPE or stored at 20° C until analysis. In order to avoid chemical decomposition of the analytes, antioxidant (BHT, 0.01% final concentration) was used from the beginning of the analytical procedure. Prior to SPE, samples were equilibrated with EDTA for 15 min, treated with isopropanol, and then vortexed for 5 min. The samples were then applied to the conditioned extraction columns, mounted in a vacuum manifold without vacuum suction, and were subjected to the extraction procedure. Eluates were collected in 1.1-ml vials (Chromacol Ltd., Welwyn Garden City, UK) and dried gently under a stream of N₂. The dried eluates were stored at 20° C for up to 24 h before further processing by HPLC.

RP-HPLC separation. Dried samples were extracted with 60 µl of mobile phase, vortexed, sonicated, centrifuged briefly, and injected into the HPLC system. As mobile phase, a 54:8:38:0.001 (vol/vol/vol/ vol) mixture of MeCN, MeOH, water, and FA, supplemented with 6 mmol/L PSA and 20 µmol/L EDTA, was used (pH 4.38, adjusted with 10% aqueous TEA). This technique achieved baseline separation of all eicosanoids within 43 min (Figure 1), except for the coeluting 20-COOH-LTB₄ and 20-HO-LTB₄, 6t-LTB₄ and 6t,12e-LTB₄, as well as 8-HETE and 12-HETE. The photodiode array detector that was used to monitor the eluting analytes provided full UV spectra (190 to 340 nm) of eluting compounds and allowed checking for peak purity and subtraction of possible coeluting material. The eicosanoids were identified and quantified at 270 nm (LTs), 237 nm (HETEs), and 204 nm (EETs), respectively, by spectra plot on the peak maximum, including use of the peak purity control method, coelution with commercial standards, and by internal standard method employing PGB₁.

Isolated, Ventilated and Perfused Human Lung

The technique of isolated lung perfusion and ventilation used in our laboratory has been previously described (21), and was adopted in the present study to whole human lungs (n = 8). These came from patients undergoing lung resection for bronchial carcinoma. This part of the study was approved by the university ethics committee of our institution, and informed consent was obtained from each patient prior to surgery. Immediately after resection, the vascular bed was extensively rinsed with Krebs-Henseleit buffer at 37° C containing 132.8 mmol/L NaCl, 4.3 mmol/L KCl, 1.1 mmol/L KH₂PO₄, 24.1 mmol/L NaHCO₃, 2.4 mmol/L CaCl₂, 1.3 mmol/L MgCl₂, and 240 mg of glucose per 100 ml; pH = 7.35 to 7.45. The area of carcinoma was clamped, and for washout of blood the perfusate was initially not recirculated. The perfusate flow was slowly increased to 400 ml/min. The lungs were then placed in a temperature-equilibrated housing chamber at 37° C, where they were freely suspended from a force transducer for monitoring of lung weight, and were ventilated with 5.3% CO₂, 21% O₂, and 73.7% N₂ (tidal volume = 500 ml; ventilatory frequency = 12 inflations/min; end-expiratory pressure, 4 mm Hg). The alternate use of two separate perfusion circuits allowed the repetitive exchange of perfusion fluid. After a steady-state period of 20 to 30 min, the perfusate was exchanged and the lungs were recirculatingly perfused at a pulsatile flow rate of 600 ml/min (500 ml total volume). After an interval of 15 min, A23187 was admixed with the perfusion fluid at a final concentration of 2 µmol/L, and time was set to zero. Perfusate samples of 10 ml each for eicosanoid analysis were collected from the venous effluent immediately before (0 min) as well as 5, 10, 20, and 30 min after ionophore challenge. Pulmonary artery pressure (Ppa) and weight gain (W) of the isolated organ were registered continuously.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Human lung liberation of a broad spectrum of AA-derived LO and monooxygenase products was noted. By means of their chromatographic and spectroscopic characteristics, we identified the following distinct groups, eluting sequentially (Figures 1 and 2): (1) conjugated trienes (LTs, DiHETEs) with specific spectral maxima (269 to 282 nm), compounds of high polarity that eluted with the described chromatographic system from 3 to 18 min, which were detected at 270 nm; (2) cis/trans-conjugated allylic monohydroxydienes (HETEs) having specific UV maxima between 233 to 238 nm, molecules of intermediate polarity, with retention times of 18 to 28 min, which were detected at 237 nm; and (3) eicosanoids without conjugated double bonds (EETs, 5,6-DHET -lactone), with optical absorption maxima at 199 to 202 nm, metabolites of low polarity, which eluted at the end of the chromatogram (28 to 44 min) and were detected at 204 nm.

View larger version (37K): [in this window] [in a new window]   Figure 2.   (Top panel ) RP-HPLC profile of eicosanoids released from human lung upon stimulation with A23187. (Bottom panels) Spectra plots of peak maxima registered simultaneously by photodiode array detector and used for identification and peak purity control of endogenous lipid mediators.

In the absence of ionophore challenge, no baseline LT release was detected. A23187 provoked the rapid appearance of LTB₄, -ox-LTB₄ (i.e., 20-HO-LTB₄ + 20-COOH-LTB₄), peptido-LTs (LTC₄, LTD₄, LTE₄) in the recirculating perfusate (Figure 3). The yield of these AA metabolites generated enzymatically via LTA₄ far surpassed the amount of nonenzymatic LTA₄ hydrolysis products (6-trans-LTB₄, 6-trans-12-epi-LTB₄, 5S,6R-DiHETE, 5S,6S-DiHETE) (> 98% versus < 2%). Within the total number of LTs, LTE₄ and -ox-LTB₄ were the predominant compounds, with the fastest rates of appearance and highest perfusate concentrations (Table 1), and exhibited ongoing release within the 30-min period after exposure to A23187. LTB₄ and LTC₄ displayed medium perfusate concentrations, with levels plateauing at ~ 10 to 20 min after ionophore challenge. LTD₄ and the nonenzymatic LTA₄ hydrolysis products appeared at the lowest concentrations.

View larger version (23K): [in this window] [in a new window]   Figure 3.   Kinetics of LT appearance in recirculating buffer fluid of isolated human lungs stimulated with the calcium ionophore A 23187 (n = 8; mean ± SEM; error bars are missing when they fall within symbol).

Some minor baseline release of 15-HETE (1.49 ± 0.08 ng/ml after 15 min) was noted (Figure 4). Upon stimulation with the calcium ionophore, human lungs secreted all six regioisomeric monohydroxydienes, with 5-HETE as the predominant and 9-HETE as the minor compound. The total sum of all HETEs was approximately 45% of the total quantity of all LTs. All HETEs displayed ongoing release over the 30-min observation period following A23187 challenge.

View larger version (22K): [in this window] [in a new window]   Figure 4.   Time course of HETE release in recirculating buffer fluid of isolated human lungs stimulated with the calcium ionophore A 23187 (n = 8; mean ± SEM; error bars are missing when they fall within symbol).

Under baseline conditions, we noted substantial liberation of 11,12-EET and 8,9-EET (28.19 ± 1.96 ng/ml and 19.68 ± 1.89 ng/ml, respectively, within 15 min), in contrast to that of 5,6-EET and 14,15-EET (Figure 5). A23187 challenge provoked dramatically enhanced and progressive 8,9-EET release: this epoxygenase product displayed by far the most rapid rate of appearance and highest maximal perfusate concentration of all eicosanoids analyzed in our study (Table 1). In addition, the ionophore induced some minor liberation of 14,15-EET and 5,6-EET, including the latter's nonenzymatic intramolecular esterification product, 5,6-DHET -lactone. In contrast, the time course of baseline 11,12-EET release was more or less unaffected by the ionophore challenge. The total quantity of the four regioisomeric EETs, including 5,6-DHET -lactone, surpassed the total sum of all LTs and HETEs by approximately twofold.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Kinetics of EET appearance in recirculating buffer fluid of isolated human lungs stimulated with the calcium ionophore A 23187 (n = 8; mean ± SEM; error bars are missing when they fall within symbol).

A near constant baseline Ppa was noted in all experiments, with values ranging from 9 to 10 mm Hg. In response to A23187 challenge, a rapid increase occurred in Ppa, with a maximum pressure increase of ~ 5 mm Hg after 5 min (Figure 6). In addition, stimulation with A23187 ionophore induced progressive lung edema formation, given as weight gain (W) in Figure 6.

View larger version (16K): [in this window] [in a new window]   Figure 6.   Increases in Ppa and W in isolated human lungs stimulated by intravascular application of A 23187 (n = 8; mean ± SEM; error bars are missing when they fall within symbol).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In accord with previous investigations done with lungs of different species (1, 17), the current study demonstrates that the human pulmonary vasculature is a rich source of LT synthesis. In response to calcium ionophore challenge, pronounced LT generation was noted, proceeding nearly exclusively via enzymatic pathways (> 98% of all products), with a contribution of < 2% of nonenzymatic LTA₄ hydrolysis products. The enzymatic conversion of LTA₄ via glutathione-S-transferase (EC 2.3.2.2) gave way to the appearance of LTC₄, which was further bioconverted via LTD₄ to the accumulating end product LTE₄. In addition, LTA₄ hydrolase activity, with the generation of LTB₄ and further catabolism to -ox-LTB₄, was a prominent finding in the human lung vasculature.

Attribution of the LT metabolic pathways to different vascular and perivascular cell types was beyond the scope of the current study; however, in concert with experimental data, some features of this may be outlined. Lungs are known to harbor a large ("resident") capillary pool of neutrophils, monocytes, and lymphocytes even after extensive rinsing with buffer fluid (22, 23), adding to the metabolic capacities of endothelial cells (EC), macrophages, vascular smooth-muscle cells (VSMC), and mast cells, whereas erythrocytes and platelets are largely washed out by perfusion. Because 5-LO is restricted to cells of the myeloid lineage (24), leukocytes and macrophages, but not EC and VSMC, must be assumed to initialize LT synthesis via LTA₄ formation. LTB₄ generation may then occur in the various leukocyte and macrophage types as well as in ECs, although the rapid appearance of the -ox products of LTB₄ suggests major involvement of neutrophilic granulocytes, since this is the only cell type equipped with substantial -oxygenase (a cytochrome P450 isoenzyme) activity. In contrast, glutathione-S-transferase, giving way to peptido-LT synthesis, is not expressed in neutrophils, but is expressed in mast cells, eosinophils, and monocyte/macrophages, as well as in ECs and VSMCs.

Previous studies with intravascular bolus injection of intact LTA₄ (4) or introduction of freshly prepared human polymorphonuclear neutrophils (5, 25) into perfused rabbit lungs, and experiments with neutrophil-EC cocultures (24, 26), showed transcellular peptido-LT synthesis, including a neutrophil-EC LTA₄ shift and subsequent bioconversion of the allylic epoxide to LTC₄ in EC. Similarly, VSMC may serve as acceptor cells for extracellularly presented LTA₄ (28). Such transcellular formation of peptido-LTs may thus contribute substantially to the strong potency of the human lung vasculature for peptido-LT synthesis in response to inflammatory challenge.

The six regioisomeric mono-HETEs are generally known to be synthesized either by the respective LOs (5-LO, 8-LO, 9-LO, 11-LO, 12-LO, and 15-LO) or by cytochrome P450 isozymes (2, 10, 29, 30), and even by cyclooxygenase (31). Interestingly, we noted some continuous liberation of 15-HETE, a ubiquitously generated eicosanoid having antiinflammatory features and previously isolated from fragments of human lung (2, 32), under baseline conditions. In response to the administration of A23187 ionophore, we noted substantial ongoing formation of the proinflammatory agent 5-HETE, which rapidly became the predominant compound within the group of HETEs. This finding corresponds to data from perfused rabbit and rat lungs subjected to inflammatory challenge (1, 2, 33). In addition, we observed the formation of 8-HETE/12-HETE, 9-HETE, and 11-HETE, the biologic properties of which are less extensively characterized. Overall, HETE generation in the human lung vasculature was less prominent than LT synthesis in this compartment.

To the best of our knowledge, the present study is the first to address the release of cytochrome P450-derived cis-EETs into the vascular space of intact lungs. Most impressively, the overall quantity of oxiranes detected in this compartment far surpassed the total amounts of LTs and HETEs, and the monooxygenase products of AA showed the fastest rates of appearance. This finding corroborates and substantially extends previous reports of EET formation in microsomal fractions or tissue of rabbit (6), canine (7), guinea pig (8), and human lung (9) in vitro. The cytochrome P450 isoenzymes implicated in EET synthesis in these fractions include CYP2B4 (6, 8), CYP2J3, and CYP2J2 (9). Analyzing the kinetics of EET appearance in the human lung perfusate, we noted three different reaction patterns: (1) 11,12-EET showed by far the highest baseline liberation of all eicosanoids that we investigated (~ 19-fold the baseline liberation of 15-HETE), the kinetics of which did not substantially change in response to ionophore challenge. This finding supports the view that EETs may represent endogenous cellular constituents (partly stored in esterified form in membrane glycerolipids) involved in maintaining homeostatic functions (6, 10). (2) 8,9-EET showed somewhat slower baseline liberation than 11,12-EET, but its liberation was dramatically enhanced in response to A23187 challenge, resulting in the highest buffer concentrations by far of any single lipid mediator that we measured (> 150 ng/ml). (3) 5,6-EET and 14,15-EET were nearly absent under baseline conditions, but responded with continuous release to ionophore provocation; in the case of the labile 5,6-EET, its metabolite 5,6-DHET -lactone was found to rapidly become the major compound. Considering in vitro findings of EET formation in cultured EC and VSMC (14, 15), and the immunohistochemical demonstration of CYP2J2 in human pulmonary vascular endothelium and the vascular smooth-muscle layer (9), these cell types are sources of the rich cis-EET formation in the human lung vasculature. In addition, monocytes/macrophages may add to the oxirane generation in this compartment.

Assessment of the physiologic or pathophysiologic functions of the different LTs, HETEs, and EETs in the human lung vasculature was beyond the scope of the present investigation. Considering their vasoconstrictor and permeability-increasing properties (2, 24), peptido-LTs may well be involved in the pressor and leakage response to A23187. On the other hand, EETs, with their regio- or stereoselective vasodilatory properties (10, 11, 34, 35), may contribute significantly to maintenance of the physiologically low baseline pulmonary artery pressure, and may limit the extent of the vasoconstrictor response. However, cyclooxygenase products elicited by ionophore challenge may also modulate pulmonary vascular tone. Interestingly, in pilot experiments, we identified 6-keto-PGF₁, the stable degradation product of PGI₂, as the predominant prostanoid in isolated human lungs, and it may also contribute to the maintenance of low pulmonary vascular resistance. The high yield of AA epoxidation products in the human lung vasculature is particularly intriguing in view of the recent suggestion that these agents may represent the chemical substrates of the endothelium-derived hyperpolarizing factor (14). Even with dilution in the buffer fluid, the concentrations of EETs measured in the present study matched those used for provoking vascular smooth-muscle hyperpolarization in vitro (15, 16). The cytochrome P450 system of the lung, until now considered to be primarily involved in the metabolism of xenobiotics, such as inhaled carcinogens, may thus turn out to be a major constitutive pathway of AA bioactivation, generating products essential for homeostatic regulation and for pathophysiologic events in the human lung vasculature.