Activation of Nitric Oxide Release and Oxidative Metabolism by Leukotrienes B₄, C₄, and D₄ in Human Polymorphonuclear Leukocytes

NITRIC OXIDE (NO) IS AN important effector and mediator of a multitude of biological functions.1 It is enzymatically formed from L-arginine by the nitric oxide synthase (NOS) family of enzymes, either constitutive (cNOS), or inducible (iNOS).2,3 The cNOS requires calcium and calmodulin for activation and produces only small amounts of NO, whereas iNOS is independent of added Ca²⁺ and produces large amounts of NO.1-3 Several lines of evidence suggest that NO plays an important role in the inflammatory process, such as being cytoprotective as well as cytotoxic, eg, in rheumatoid arthritis and vasculitides.4,5 Furthermore, increased NO production in the airways has been demonstrated in allergic inflammatory conditions in humans,6 in disorders with increased plasma leakage,7 and for functions in the control of the proinflammatory cascade systems in allergy.8 

A cNOS has been purified from human polymorphonuclear granulocytes (PMN).9 In patients with urinary tract infections, PMN have been shown to express both iNOS and cNOS mRNA, and iNOS protein and activity was demonstrated.10 Still, the mechanisms for activation of NO production in human PMN as well as its biological significance are largely unknown. We have previously shown that human PMN produce NO rapidly on activation with bacterial oligopeptides or with a phorbol ester.11-13 

Thus, we asked whether proinflammatory leukotrienes (LT) affect NO generation in PMN. Leukotriene B₄ (LTB₄) is a potent activator of PMN migration and chemotaxis,14,15 but it also stimulates the secretion of granule enzymes16 and superoxide ions in PMN.17-19 Furthermore, others have shown that LTB₄ may activate nitrite production, conceivably dependent on NO formation, in human PMN.20 Contrary to LTB₄, the cysteinyl leukotrienes (LTC₄ and LTD₄) do not elicit adhesive or secretory responses in PMN, but are powerful mediators in allergy and inflammation by induction of plasma leakage and reduction of myocardial contractility,21effects also reported for NO.7,22 Human PMN have been reported to possess receptors for LTC₄,23,24and LTD₄ may induce elevation of intracellular Ca²⁺ in PMN.25 However, whether LTC₄ and LTD₄ affect NO production from PMN is still unknown.

Based on these considerations, we have studied the effects of LTB₄, LTC₄, and LTD₄ on NO release from human PMN. Because interaction between NO and the products of the oxidative metabolism in PMN, primarily the release of superoxide anions, may be of significance,26 we have also compared effects on NO production with the effects on superoxide release. For detection of these products, we have used highly sensitive, real-time methods.11-13,27-30 

Percoll and Sephadex G25 were from Pharmacia Fine Chemicals (Uppsala, Sweden). Polymonoprep was from Nycomed Pharma AS (Oslo, Norway). Endotoxin-free, deionized water and Hanks’ balanced salt solutions (HBSS), with or without Ca²⁺, were from GIBCO (Paisley, Scotland). Tetrakis (3-methoxy-4-hydroxyphenyl) nickel(II)porphyrin was from Interchim (Montluçon, France), Nafion and pure NO gas from Aldrich (Milwaukee, WI). L-arginine, catalase, cytochrome C (horse heart type III), 1,2-bis(2-aminophenoxy)-ethane-N,N,N’,N’-tetraacetic acid (BAPTA-AM), luminol, N-formyl-methionyl-leucyl-phenylalanine (fMLP), sulfanilic acid, N-(1-naphtyl)ethylenediamine dihydrochloride, cytochalasin B, Bordetella pertussis toxin, and bovine hemoglobin were from Sigma Chemical Co (St Louis, MO). 5S-HETE, 5-oxo-ETE, and leukotrienes B₄, C₄, and D₄ were from Cascade Biochemical (Berkshire, UK). 5S,12S-diHETE was from Biomol (Plymouth Meeting, PA). Superoxide dismutase (SOD) was from Boehringer Mannheim (Mannheim, Germany). N^G-monomethyl-L-arginine (L-NMMA) was from Calbiochem (La Jolla, CA). SK&F 104 353 was a kind gift from Prof S.-E. Dahlén, Department of Physiology, the Karolinska Institute (Stockholm, Sweden), and CP-105,696 was a kind gift from Central Research Division, Pfizer Inc (Groton, CT).

Preparation of PMN was performed by Percoll density centrifugation, essentially as described previously.27 Only freshly prepared, nominally endotoxin-free solutions were used, and all preparations were performed under aseptic conditions. The cells were resuspended in HBSS with Ca²⁺ to 3.5 × 10⁶ PMN/mL and kept at 4°C before analysis. The cells were then pelleted, washed three times in HBSS with or without Ca²⁺, and finally suspended to desired final concentration. Supernatants from the final wash were always assessed spectrophotometrically for residual hemoglobin, and the PMN used only if no such residues were found. The PMN preparation contained 98% to 99% granulocytes; the contaminating cells (1% to 2%) were principally monocytes and 0.1% to 0.7% eosinophils. To assess if this small monocyte contamination contributed to the NO produced, these cells were purified to 98% purity on Polymonoprep as described previously.31 In three separate experiments, performed in duplicate, no NO production was recorded (using the HbO₂method; vide infra) from monocytes (10⁵ cells per mL) either unstimulated or activated with fMLP, LTB₄, or LTC₄ (all at 100 nmol/L). Thus, we conclude that the contaminating monocytes do not add to the NO measured from PMN.

The assay was performed as described previously.11-13Briefly, a solution of bovine hemoglobin was made in doubly distilled, deionized water at a concentration of 1 mmol/L and oxygenated by bubbling with O₂ for 5 minutes. Subsequently, the hemoglobin was reduced with 1.5 mmol/L deoxygenated sodium dithionite in distilled water, and reoxygenated for 20 minutes. The sodium dithionite was then removed on a Sephadex G25 column. Oxyhemoglobin was prepared freshly for each experimental day and quality tested by scanning at 400 to 600 nm. The viability of the cells under study, assessed with trypan blue exclusion, was not affected by the used concentrations of, or by incubation times with, HbO₂ (5 μmol/L). Neither did used concentrations of L-arginine, L-NMMA, nor incubation in Ca²⁺-free HBSS, with or without BAPTA-AM (5 μmol/L), decrease viability.

PMN (1.75 × 10⁶/mL in HBSS with or without Ca²⁺, as indicated), or monocytes (10⁵ cells per mL) in control experiments, were treated with either 300 μmol/L L-arginine or 1 mmol/L L-NMMA for 45 minutes at 37°C in a shaking water bath. Before incubation, SOD (150 U/mL), and catalase (300 U/mL), and in certain experiments, Bordetella pertussis toxin (PT; 750 ng/mL), were added. In experiments with the Ca²⁺-chelator BAPTA-AM (5 μmol/L), this was added 15 minutes before stimulation. HbO₂ and stimuli, or corresponding volumes of HBSS, were added immediately before analysis.11 Cytochalasin B (5 μg/mL), when used, was added 3 minutes before stimulation. The metHb generation was followed for 30 minutes at 37°C, as indicated, by spectroscopy at 401 versus 411 nm in a Perkin-Elmer Lamda 7 spectrophotometer, essentially as described previously.11All samples were assessed as the difference between samples with L-arginine and identical samples with L-NMMA instead of L-arginine. Thus, only the L-NMMA inhibitable response was assessed and taken to represent the net amount of NO released from PMN. An ε = 19.7 mmol/L^-1 cm^-1 was used for calculating produced amount of NO.32 

This was performed with a three-electrode potentiostatic Biopulse system (Tacussel, Lyon, France). The working electrode was a carbon fiber (8 μm diameter, approximately 1 mm length), coated with tetrakis(3-methoxy-4-hydroxyphenyl)nickel (II)porphyrin and Nafion films29,30 and the measurement of NO production was performed as described previously.29,30 The suspension of PMN (1.75 × 10⁶/mL in HBSS with Ca²⁺) was incubated for 20 minutes at 37°C with L-arginine or L-NMMA (both at 1 mmol/L) and SOD (150 U/mL). Leukotriene B₄ and C₄ were added when a stable current baseline was reached, and the NO production was followed for 5 minutes. There were no changes in baseline current when L-NMMA, L-arginine, or SOD were added to unstimulated PMN, neither for different concentrations of H₂O₂ nor on preincubation with catalase (data not shown). The electrode was calibrated with standard NO solutions as described previously,29,30 and a standard curve with PMN present was made for each experiment and at the end of each measurement. All samples were assessed as the difference between samples with L-arginine and identical samples with L-NMMA instead of L-arginine. Thus, only the L-NMMA inhibitable response was assessed and taken to represent the net amount of NO released from PMN.

This was performed as previously described.27 Briefly, PMN at 1.75 × 10⁶/mL in HBSS, with 100 μmol/L cytochrome C, were activated and the reduction of cytochrome C was followed continuously at 550 nm in a Perkin Elmer Lamda 7 spectrophotometer with a thermostated multicuvette holder. Blanks were identical to samples, but with SOD present (150 U/mL). Thus, only the SOD-inhibitable reduction of cytochrome-C was measured in all instances. In experiments with L-arginine or L-NMMA, PMN were treated for 45 minutes at 37°C in a shaking water bath with arginines before O₂^--measurement. When used, as indicated, BAPTA-AM (5 μmol/L) was added 15 min before stimulation. An ε = 21.1 mmol/L^-1 cm^-1 was used to calculate the amount of O₂^--production.27,28 

CL was performed essentially as described previously27,28in a luminoaggregometer (Chronolog Havertown, PA). The CL response was measured from the magnitude of the peak response, given in mV, corrected for magnitude of the background (unstimulated) chemiluminescence. The reading was continuous during the luminescent reaction. All recordings were made from 0.5 or 1.0 mL samples to which reagents (5 to 20 μL) were added. Samples containing neutrophils (1.25 × 10⁶ cells/mL) were adjusted to 37°C for 3 minutes before addition of stimulus.

Statistical assessment was performed with Student’s t-test.

When PMN were activated with LTB₄ (100 nmol/L), LTC₄ (100 nmol/L) or LTD₄ (100 nmol/L), they responded with a methemoglobin production in a stimulus-dependent manner, but less pronounced compared with a structurally unrelated activator of chemotaxis and oxidative metabolism, the oligopeptide fMLP (100 nmol/L) (Fig 1). The release of NO was dose-dependent for the three tested leukotrienes (Fig 2) and the rank order of potency in inducing NO release was LTB₄ = LTC₄ > LTD₄ (Fig 2). The LTB₄ response was completed after approximately 15 minutes, whereas the fMLP-stimulated PMN still showed NO production at this time.

Next, we assessed if the detection of NO-release from PMN with the HbO₂-method was as specific for NO when PMN were activated with leukotrienes, as previously shown when a formylpeptide or phorbolester were used as stimuli.11-13 To this end, a highly sensitive and specific porphyrinic microsensor was used.29,30 With LTB₄ (100 nmol/L), a prompt and continuous release of NO from PMN was recorded (Fig 3). Production of NO 5 minutes after addition of LTB₄ was 18.7 ± 4.5 pmol NO/10⁶PMN (n = 4). This was in good agreement with the results obtained with the HbO₂ method (12.2 ± 4.5 pmol NO/10⁶ PMN and minutes; n = 8). Leukotrienes C₄ or D₄ (at 100 nmol/L) also conferred NO generation, measurable with the electrode, in a similar concentration as obtained with the HbO₂ method (13.5 ± 5.5 and 5.65 ± 1.5 pmol NO/10⁶ PMN, respectively; n = 2). Time to response was similar for all three leukotrienes with a lag period of 30 seconds. Because the HbO₂ method permits kinetic studies also for longer time periods than the NO electrode, this method was selected for further studies.

Next, we assessed if the LTB₄-induced NO formation was the result of a stereospecific, ie, receptor-mediated process, or an unspecified effect on PMN. To this end, we applied related eicosanoids, 5-hydroxy eicosatetraenoic acid (5S-HETE; 100 nmol/L), or its metabolite 5-oxo-ETE (100 nmol/L), or 5S,12S-diHETE (10 and 100 nmol/L),33,34which have been shown not to activate the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase or chemotaxis in PMN. However, none of these eicosanoids showed any statistically significant release of NO from PMN (28 ± 10, 29 ± 11, 24 ± 13 and 0 pmol NO/10⁶ PMN, respectively, 30 minutes after stimulation; mean ± standard error of mean [SEM]; n = 3). As reported previously, lipoxin A₄ did not elicit NO production from human PMN, measured with a similar HbO₂-technique as used here.5 

To further study the receptor-dependence of the leukotriene-induced responses, we applied specific receptor antagonists. Treatment of PMN with the cysteinyl leukotriene receptor-antagonist SK&F 104 35323 significantly reduced the NO-release induced by LTC₄ or LTD₄(Table 1). Substituting the cysteinyl leukotrienes with LTB₄ as stimulus assessed the specificity of this antagonist. However, no inhibition of LTB₄-induced NO formation or CL by SK&F 104 353-treated PMN was found. Conversely, treatment of PMN with the LTB₄-receptor antagonist CP-105,69635 significantly reduced NO production in LTB₄-stimulated PMN (Table 1), whereas the responses to LTC₄ or LTD₄ were unaffected. Also, the effect of CP-105,696 on oxidative metabolism was assessed with CL. Because LTC₄ and LTD₄ did not induce a CL response from PMN, we used fMLP-activation of PMN as positive control. The CL response from CP-105,696-treated PMN was totally abolished on LTB₄-activation (49.6 ± 9.9 mV without CP-105,696 compared with 0 mV with CP-105,696, mean ± SEM; n = 4), but no reduction of the response to fMLP was found (635 ± 19.6 and 650 ± 53.5 mV, with and without CP-105,696, respectively; mean ± SEM; n = 6). We have previously shown that SK&F 104 353 inhibits the rise of cytosolic calcium concentration in LTC₄- and LTD₄-stimulated PMN.24 

To further establish a role for serpentine receptor transduction mechanisms, ie, PT-sensitive G-proteins, PT (750 ng/mL) was added during incubation. We found a marked blunting of NO production in PT-treated PMN for all three stimuli, LTB₄, LTC₄, and LTD₄ (at 100 nmol/L), the reduction being 91.2% ± 4.4%, 96.7% ± 10.4%, and 98.0% ± 5.1%, respectively (mean ± SEM; n = 4; P < .05) 15 minutes after stimulation compared with simultaneously run untreated controls. Likewise, PT treatment reduced fMLP- and LTB₄-stimulated CL with 85% ± 4.5% (mean ± SEM; n = 3; P < .05). Consequently, these results support the concept that NO production in PMN, activated by these leukotrienes, is a cell surface receptor-dependent event involving receptors with specificity for LTB₄ or the cysteinyl leukotrienes, respectively.

We next asked if NO release was dependent on cytosolic Ca²⁺-transients distal to G-protein activation, as previously described for LTB₄ activation of other PMN functions.16,24 To this end, PMN were treated with the high-affinity cytosolic Ca²⁺-chelator BAPTA-AM (5 μmol/L), and subsequently suspended in nominally calcium-free HBSS. These conditions markedly reduced the NO release induced by LTB₄ and LTD₄, and to a significant, but slightly less extent, by LTC₄(Fig 4). The same experimental conditions were used to assess NADPH oxidase activity induced by LTB₄or fMLP, by means of CL or cytochrome-C reduction. We found that treatment with BAPTA-AM reduced the responses significantly in both assays. Addition of BAPTA-AM reduced LTB₄ induced CL from 49.6 ± 9.9 to 7.5 ± 1.7 mV and for fMLP from 650 ± 53.0 to 15 ± 2.5 mV (mean ± SEM; n = 3 to 4; P< .05), in accordance with previous reports.18,25 The reduction of O₂ production assessed with cytochrome-C reduction was in the same range; in fMLP-stimulated PMN, the production was reduced from 10.45 ± 0.9 to 1.7 ± 0.4 nmol O₂^-/10⁶ PMN and for LTB₄ from 2.0 ± 0.4 to 0 nmol O₂^-/10⁶ PMN, 10 minutes after stimulation (mean ± SEM; n = 8). Neither LTC₄ nor LTD₄ induced a CL response or cytochrome-C reduction from PMN under normocalcemic conditions. To assess the possibility that this might be due to concomitant NO production and scavenging of O₂^-, we also assessed BAPTA-AM-treated PMN by CL after stimulation with LTC₄ or LTD₄. However, similar to nontreated PMN, the cysteinyl leukotrienes did not evoke a CL response from PMN treated with BAPTA-AM, irrespective of the Ca²⁺ content in the medium. Neither did the addition of BAPTA-AM during incubation reduce the PMA-induced CL response in human PMN (1505 ± 53 and 1415 ± 48 mV, respectively; mean ± SEM; n = 2).

Finally, we studied the dependence of the cytoskeleton and degranulation for the NO-response in LTB₄-stimulated PMN treated with cytochalasin B, previously reported to enhance the chemiluminescence responses in LTB₄-stimulated PMN and PMN aggregation.18 Both the initial and the sustained NO release was significantly increased in samples preincubated with cytochalasin B (Fig 5), and the mean increase in NO production during the observation period was 86% ± 32% (mean ± SEM; n = 3; P < .05 compared with identical samples without cytochalasin B).

The production of NO by human PMN has previously been demonstrated by us (using the same methods),11-13 as well as by others (using various methods for detection of NOS activity).10,20,36 Because NO is highly reactive and has a very short half-life in biological systems,32 its detection in phagocytes is technically demanding. We have previously reported the characteristics of the two methods.11-13,28,29 Thus, both methods are assessing NO directly and with considerably higher sensitivity than, for example, colorimetric analysis of nitrite. Moreover, one might speculate that superoxide anions could attenuate the NO assessments. However, previous studies in PMN from patients with chronic granulomatous disease show that these PMN, incapable of superoxide anion production, oxidized HbO₂ to approximately the same extent as normal PMN.11,37 

In previous studies, we17,18 and others25,38have demonstrated that functional responses of PMN, eg, activation of NADPH oxidase, homotypic aggregation, and [Ca²⁺]_i transients elicited by LTB₄ are more rapid in onset, as well as terminated earlier than those evoked by fMLP and, particularly, the phorbol ester PMA. The reason remains unclear, but was attributed to kinetic differences in stimulus-response coupling systems. Here, we show that a similar difference in activation kinetics for NOS exists, where leukotriene responses reached a plateau phase earlier than the fMLP responses (and were more rapid in onset than NO generation evoked by PMA).18,28 These findings point to consistent and stimulus-specific activation pathways that apply also for the NOS studied here. Furthermore, LTC₄ or LTD₄activate the production of NO in sharp contrast to their lack of activity in inducing NADPH-oxidase activation,17,25 and with kinetic properties similar to LTB₄. There was also a difference between the NADPH oxidase and NOS in activation kinetics, in that all NO responses were slower in onset. Because our detection systems give instant recordings of NO production, we believe that this NOS requires more time to produce NO than the NADPH oxidase to assemble and generate superoxide ions.

The concept that the leukotriene-induced NO release was mediated by cell surface receptors was supported by the results of cysteinyl-leukotrienes and LTB₄-receptor antagonists. These results are similar to findings for a variety of other responses to leukotrienes in PMN and other cells. Moreover, the inhibitory effect of PT, indicating that the NO response was transduced by means of PT-inhibitable G-proteins, is similar to the activation of the NADPH oxidase. It has previously been demonstrated that the cytoskeleton and actin levels are important for receptor expression in PMN.35 Depolymerization of the actin cytoskeleton with cytochalasin B enhanced the LTB₄-induced NO response in the same manner as for superoxide production (and homotypic aggregation). These results show that the events leading to the activation of the NOS studied here are highly regulated and follow similar principles as aggregation and superoxide forming reactions to LTB₄ and fMLP.

To assess a key step of the signal transduction in PMN, we studied the significance of cytosolic Ca²⁺, [Ca²⁺]_i transients for the leukotriene-induced NO release. Indeed, the leukotriene-induced NO release showed a similar dependence on [Ca²⁺]_i as that for fMLP-induced NADPH oxidase activation.25 Previous studies have found that LTB₄ is a substantially more potent activator of [Ca²⁺]_i release in human PMN than LTC₄ or LTD₄.24,25 That finding is still consistent with our data showing that the NO release activated by cysteinyl leukotrienes has a similar requirement for access to cytosolic Ca²⁺ as that in response to LTB₄. However, the nature of the NOS activated in these cells, constitutive or inducible, is not possible to predict from the present data.

Thus, in the present study, we have demonstrated that LTB₄activates human PMN to release NO in a highly regulated receptor-dependent process involving cytosolic Ca²⁺transients, in apparent analogy with other secretory events induced in PMN by LTB₄, such as superoxide anion production.18,25,28 However, the cysteinyl-leukotrienes, LTC₄ and LTD₄, act in a similar manner as LTB₄ and to virtually the same extent to activate NO release from PMN, but do not induce a measurable activation of the NADPH oxidase, as found here and in previous studies by others.25 This is, to the best of our knowledge, the first demonstration of an effect of cysteinyl-leukotrienes for secretory events in PMN. Consequently, the human PMN seems capable to tailor its production of radicals, NO, or O₂^-, in a stimulus-dependent fashion. The mechanism for this selective response pattern remains to be elucidated. Another question that is raised is the possible biological importance of these findings. Although the amount of NO produced in each granulocyte is very low, the total number of cells and their tendency to migrate towards inflammatory and infectious foci could imply that a NO production is induced in affected tissues and that it is important for the functional responses in human PMN.

The authors wish to thank Anette Landström, Annie Brunet, and Hussein Fazipour for skillful technical assistance.