Nitrous oxide has been shown inconsistently to impair the oxidative function of neutrophils. The choice of the stimulus, receptor agonists, or stimuli acting independent of receptors seems to determine whether nitrous oxide impairs the oxidative functions, suggesting an interference with the cytosolic signaling of neutrophils.
Production of hydrogen peroxide by neutrophils was assessed using flow cytometric analysis. N-formyl-methionyl-leucyl-phenylalanine (FMLP), C5a, dioctanylglycerol, and phorbol-12-myristate-13-acetate were used as stimuli. In addition, the expression of receptors for FMLP and the cytosolic-free calcium response of cells were measured.
Nitrous oxide depresses C5a- or FMLP-induced generation of reactive oxygen derivatives in a concentration-dependent manner. However, the response with direct activation of protein kinase C was unaffected. Further, the number of FMLP receptors and the cytosolic calcium response were unaffected. Inhibition of the oxidative response was not reversible within the observation period of 4 h.
Nitrous oxide inhibited the intracellular signaling of the investigated G-protein-coupled receptors for chemotactic peptides. No interference of nitrous oxide with reduced nicotinamide adenine dinucleotide phosphate oxidase, the oxidative enzyme system of neutrophils, nor with its activation through protein kinase C was detected.
POLYMORPHONUCLEAR neutrophils are a major component of the nonspecific cell-mediated immune response. A crucial mechanism of defense is the generation of reactive oxygen derivatives.
Data on the effects of nitrous oxide (N2O) on the motility and microbicidal oxidative activity of neutrophils are controversial. Some authors, using receptor agonists, [2,3]observed the inhibitory effects of N2O, whereas others, using intracellularly acting stimuli, [4,5]have not. Presumably, the choice of the stimulus used to activate neutrophils determines whether N (2) O exerts an inhibitory effect. This suggests an interference of N (2) O with neutrophil signal reception or transduction.
To characterize the mechanism of interference, this in vitro study investigated the effect of 30% and 70% N2O on the generation of reactive oxygen derivatives in human neutrophils with respect to the signaling pathways used. The flow cytometric assay used for this study allows the quantification of the oxidative response to stimulation of multiple neutrophils at the cellular level. 
To localize potential sites of the N2O interactions, several stimuli were used to initiate neutrophil responses. The bacterial peptide N-formyl-methionyl-leucyl-phenylalanine (FMLP) and a product of the complement cascade, C5a, both potent chemoattractants for neutrophils, were used to induce hydrogen peroxide (H2O2) generation by neutrophils. Both substances are agonists of specific G-protein-coupled receptors and share the same second messenger pathway (Figure 1). Binding of a fluorescent receptor ligand to FMLP receptors was used to analyze changes in receptor numbers. The increase in cytosolic-free Ca2+[Ca2+]i, after stimulation with FMLP or C5a, was assessed for further analysis of receptor function. Finally, the receptors and parts of the second messenger cascade were bypassed through induction of an oxidative response by membrane-permeable dioctanylglycerol (DiC8-DAG) and phorbol-13-myristate-12-acetate (PMA). [9,10]Both DiC8-DAG and PMA are synthetic analogs of diacylglycerol (DAG), a compound of the signaling pathway, and induce the oxidative response by direct activation of protein kinase C, an enzyme in the neutrophil signaling chain downstream of receptors and second messengers but upstream of the enzyme system responsible for the generation of reactive oxygen products. 
Materials and Methods
This study was approved by the local ethical board of the University of Regensburg Medical School. After informed consent, venous blood was drawn from healthy donors with no history of infections 2 weeks before the experiments. The mean age of the donors (n = 6) was 32 yr, with a range of 27–38 yr. Average leukocyte counts were 6,300 +/- 1,200/micro liter (mean +/- SD), and hemoglobin levels were 15.1 +/- 1.3 g/dl. The differential leukocyte counts were as follows: neutrophils, 53 +/- 8%; lymphocytes, 32 +/- 7%; monocytes, 8 +/- 3%; eosinophils, 3 +/- 2%; and basophils, 1 +/- 1%. The complete blood count and the differential count were determined before each experiment using a Technicon H*3 counter (Bayer, Tarrytown, NY). No donors had to be excluded because of abnormal blood counts.
Leukocytes were isolated by a sedimentation of erythrocytes on Ficoll-Paque (Pharmacia, Upsalla, Sweden). Heparinized (10 U/ml) whole blood (3 ml) was layered on 3 ml lymphocyte separation medium (Ficoll; density, 1.077 g/ml). Erythrocytes aggregated at the interface and settled at room temperature without centrifugation. After 40 min, the upper 800 micro liter of the supernatant leukocyte-rich plasma was withdrawn, with care taken to avoid contact with that part of the plasma close to the interface with the separation medium. To avoid artifactual activation of cells, the isolation process did not involve lysis, centrifugation, or washing procedures.
Intracellular Hydrogen Peroxide Production
Nitrous oxide (Linde, Unterschleissheim, Germany) was tested in vitro with neutrophils from each of the six donors. The oxygen content of fresh gas was maintained at 30% in all samples (control, 30 vol%, and 70 vol%). The assay for the intracellular H2O2production was done in an airtight glovebox with a volume of about 75 1. Fresh gas was prepared using a gas-mixing unit (Draeger, Lubeck, Germany). Fresh gas flow, after equilibration of the atmosphere inside the box using high flow rates, was 5 l/min during the experiments. Inside the glovebox, the samples were kept at 37 [degree sign] Celsius. To keep the amount of evaporated water from the samples as low as possible, the tubes were sealed with gas-permeable parafilm during incubation. Nitrous oxide delivery under the hood was monitored continually using a multigas-analyzer (Capnomac Ultima; Datex, Helsinki, Finland).
Before the propylene-tubes were filled and the leukocytes added, Dulbecco's PBS (Gibco-Life Technologies, Heidelberg, Germany) was equilibrated with N2Oinside the glovebox by bubbling fresh gas through for 30 min at 37 [degree sign] Celsius. The supernatant leukocyte-rich plasma was suspended 1:50 (20 micro liter leukocytes and 980 micro liter PBS) in pre-equilibrated Dulbecco's PBS and incubated for 20 min at 37 [degree sign] Celsius. The leukocytes were loaded in the presence of N2O with the fluorogenic substrates dihydrorhodamine 123 and carboxy-seminaphthorhodafluor-1-acetoxymethylester (SNARF1/AM) for 10 min (both dyes from Molecular Probes, Eugene, OR). The overall exposure time to N2O before stimulation was 30 min. The final concentrations were 1 micro Meter for dihydrorhodamine 123 and 0.1 micro Meter for SNARF1/AM. Next FMLP, C5a, DiC8-DAG, or PMA, respectively, were added to stimulate the H2O2production of neutrophils (all from Sigma Chemical Co., Deisenhofen, Germany). After 15 min of incubation at 37 [degree sign] Celsius, the reaction was stopped on ice. Dead cells were counterstained with propidium iodide (PI; Serva, Heidelberg, Germany) at a final concentration of 30 micro Meter. The specimens were stored on ice in complete darkness and measured within 1 h. To evaluate the recovery of neutrophil function, samples of unexposed and exposed cells (70% N2O) were allowed to recover for varying periods of time before initial stimulation: during exposure, 1 h, 2 h, and finally 4 h after the end of exposure.
For analysis, we used a FACScan flow cytometer (Becton Dickinson, San Jose, CA) with argon ion laser excitation at 488 nm, with 10,000 cells of each stained sample measured. Data were acquired and processed using LYSIS-II software. After calibration with standard dye beads (Quantum 26; Flow Cytometry Standards Europe, Leiden, The Netherlands), the results of the cellular fluorescence study were expressed in molecule equivalents of soluble fluorochrome. These units were used for the absolute quantification of cellular fluorescence and allowed interassay and interlaboratory comparison of data. The dead cells were assessed by their lack of esterase activity and their propidium iodide fluorescence (> 600 nm). Leukocyte esterase activity was determined based on SNARF1-related orange fluorescence. The SNARF1/AM is cleaved in vital leukocytes by esterases to SNARF1. Neutrophils were identified by their typical side-scatter and forward scatter light patterns as well as their esterase activity. Side scatter depends primarily on the granularity of cells, whereas forward scatter is related to cell size.
Expression of Receptors for FMLP
We used the assay established by Allen et al. without modification. Briefly, leukocyte-rich plasma suspended in Dulbecco's PBS (1:50) was incubated for 30 min in the presence of N2O at 37 [degree sign]C under the same conditions as described for the H (2) O2assay. Subsequently, the samples were cooled to 4 [degree sign]C to avoid further activation of neutrophils. For staining of FMLP receptors, fluorescein-labeled formyl-Nle-Leu-Phe-Nle-Tyr-Lys was added in a final concentration of 100 nm, incubated for 10 min, and washed twice. Samples were kept in the N2O atmosphere during the staining and measuring procedure. Formyl-Nle-Leu-Phe-Nle-Tyr-Lys, a highly specific binding analog of FMLP, emitting light at 520 nm when excited at 488 nm (Molecular Probes). Neutrophils were identified based on side scatter. Nonspecific binding of formyl-Nle-Leu-Phe-Nle-Tyr-Lys was determined by adding excess FMLP (10-4M) in a parallel set of tubes.
Increase in Cytosolic-free Calcium
The increase in [Ca2+]iwas assessed using flow cytometric analysis combined with the fluorescent calcium indicator fluo-3/acetoxymethylester (fluo3/AM; Molecular Probes). The fluorescence of this 488-nm excitable indicator is enhanced in the presence of Ca2+. Leukocytes were suspended in N2O-equilibrated Dulbecco's PBS containing 2 micro Meter fluo-3/AM and incubated for 25 min in the glovebox described previously. Separate samples were prepared for controls and different concentrations of N2O. Thereafter, the tubes were sealed with gas-tight tape and immediately connected to the flow cytometer. After 60 s of recording, the stimuli, FMLP (100 nM), or C5a (200 pM) were added, and the measurement was continued for 240 s. Subsequently, ionomycin (2 micro Meter) was added. Ionomycin (Calbiochem, Bad Soden, Germany), a calcium ionophore, was used to saturate the intracellular indicator with Ca2+to standardize the analysis, independent of dye loading. The resulting increased fluorescence was recorded for 1 min.
Gas Chromatography and Mass Spectrometry
The volatility of the investigated agent may make the maintenance of stable, reproducible drug concentrations in the fluid phase of the experiment difficult. In a separate set of experiments, concentrations of N2O in the fluid and gas phases were analyzed by gas chromatography and high-resolution mass spectrometry, using as gas chromatograph the HRGC Mega 2 (Fisons Instruments SpA, Rodano, Italy). The molecular ion of N2O was recorded by single-ion monitoring, using the molecular ion of carbon dioxide as a reference mass; the cycle time was 0.2 s. Quantification was based on signal areas in the mass chromatograms. The gas chromatography and high-resolution mass spectrometry measurements showed that the equilibration with N2O of the buffer solution was complete after 5 min with a buffer/gas partition coefficient of 0.423 and a concentration of 13.4 +/- 1.2 mM in the fluid with 70% N2O at 37 [degree sign] Celsius. The concentrations of N2O during experiments were stable. Ten minutes after withdrawal of N2O, the concentration decreased to < 2% of the initial values.
All experimental data are presented as mean values with standard deviations. Lilliefors' test was used to examine normal distribution, Levene's test to check homogeneity of variance, and Student's t test to check differences between N2O-exposed and control samples assessed in parallel. To test differences between different N2O concentrations, analysis of variance and the Tukey B test were computed. Statistical significance was assumed when probability values were < 0.05.
Hydrogen Peroxide Production of Neutrophils
The small amounts of intracellular H2O2generated after stimulation could be assessed by quantifying the intracellular oxidation of the indicator dye dihydrorhodamine 123 to rhodamine 123. The former is a nonfluorescent and membrane-permeable fluorogenic substrate, whereas the latter oxidation product emits a green light (510–530 nm) on excitation. This method has several advantages over those previously used to quantify neutrophil oxidative function. Purification of neutrophils, which often leads to artifactual activation of cells, is unnecessary. Flow cytometric results are not affected by variations in the concentration of neutrophils in the assay because the oxidative response is analyzed on the single-cell level. Flow cytometric analysis revealed a heterogeneous oxidative response after receptor-dependent stimulation with either FMLP (100 nM) or C5a (200 pM). Only approximately one half of all neutrophils produced reactive oxygen products, such as H2O2, at this degree of stimulation (Figure 2, Table 1).
In the presence of N2O, the oxidative response to either FMLP or C5a stimulation decreased in a concentration-dependent manner (Figure 3and Table 1). This effect was characterized by a dose-dependent reduction in the number of neutrophils recruited for the oxidative response, in close correlation with a decrease of the total oxidative activity (Figure 3, Table 1).
The protein kinase C (PKC) activators, DiC8-DAG and PMA, were used to bypass the receptor and the downstream intracellular signaling. Both substances elicit the H2O2production through a direct activation of PKC. The reaction of neutrophils in response to both stimuli was homogeneous. DiC8-DAG (7.5 micro Meter) induced a mean oxidative response intensity comparable with that of either FMLP (100 nM) or C5a (200 pM), whereas PMA produced a much higher response. The presence of N2O had no effect on the oxidative response after direct activation of PKC with either DiC8-DAG or PMA. The percentage of dead cells was always less than 0.1% independent of drug exposure or stimulation. Less than 5% of neutrophils in control samples displayed a spontaneous oxidative response, indicating a low degree of artifactual activation of neutrophils by the leukocyte preparation and by the assay procedures.
Expression of Receptors for FMLP
To exclude changes in receptor number, the FMLP receptors of neutrophils were stained in the presence of N2O using a fluorescent ligand. The exposure time of neutrophils to N2O for these experiments was similar to that of the H2O2assays, and N2O was present during the staining procedure to exclude potential competition of N2O and the fluorescent receptor ligand, formyl-Nle-Leu-Phe-Nle-Tyr-Lys, at the binding site of the FMLP receptor. Under these conditions, N2O had no effect on the expression of FMLP receptors on the cell surface. Neutrophils exposed to N2O expressed 39,200 +/- 3,200 receptors (n = 6) for FMLP per cell compared with 37,500 +/- 3,300 receptors (n = 6) on neutrophils of the control samples, indicating no significant interference of N2O with receptor expression.
Cytosolic-free Calcium Increase
Early steps of signaling, leading to the induction of the oxidative response, were further analyzed based on the increase of cytosolic-free calcium [Ca2+]ion stimulation. Both tested receptor agonists, C5a (200 pM) and FMLP (100 nM), induced a similar [Ca (2+)]iresponse (Table 2). Cytosolic-free Ca2+was maximal after 30.3 +/- 3.1 s of stimulation in N2O exposed cells and after 33 +/- 2.7 s in control neutrophils. The time course of this reactiont was comparable with and without N2O. Peak concentrations of calcium were similar in both groups (Table 2). Direct activation of PKC was not associated with an increase in cytosolic-free calcium (data not shown).
Reversibility of Functional Changes
The recovery of the oxidative response to FMLP was analyzed further. For each time point, a separate sample was prepared to avoid sequential stimulations of cells. The impairment of H2O2generation with FMLP stimulation could not be reversed within an observation time of 4 h (Table 3). The increased ability of neutrophils to generate H2O2during the 4 h of observation was considered an effect correlating with the spontaneous activation of neutrophils on prolonged storage because it was visible in N2O-exposed and control samples. Nevertheless, the extent of suppression after N2O exposure was similar at all times.
Some studies, in contrast to others, have reported an impairment of the neutrophil function, especially of the oxidative function, in the presence of N2O. [2–5]However, neither the mechanism nor the potential site of this N2O interaction are known. The purpose of this study was to investigate the influence of N (2) O on the oxidative response of neutrophils with respect to immunologically relevant, G-protein-coupled receptors, and the subsequent intracellular signaling.
Signaling Pathway of FMLP and C5a
The bacterial peptide FMLP is a physiologic agonist for a specific receptor on the surface of neutrophils. C5a, a product of the complement cascade, has been shown to act as a major chemoattractant for neutrophils, similarly to FMLP. Further, phagocytosis of Escherichia coli is mainly mediated by FMLP and its receptor on the neutrophil surface. FMLP and C5a trigger multiple neutrophil responses, including exocytosis of enzymes and superoxide generation by the reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase complex. Binding of FMLP or C5a to their G-protein-coupled receptors induces a transient elevation of inositol-1,4,5-trisphosphate, free cytosolic calcium [Ca2+]i, and DAG [10,16](Figure 1). The DAG accumulation is biphasic: The initial phase is catalyzed by phospholipase C and occurs in parallel to the generation of inositol-1,4,5-trisphosphate, whereas the second phase is more sustained and of greater magnitude. The DAG, accumulating in the delayed phase, derives from phosphatylcholine via sequential involvement of phospholipase D, which produces phosphatidic acid (PA) and phosphatidate phosphohydrolase. [18,19]The accumulation of PA and DAG are supposed to mediate sustained cellular responses. [7,20]As shown in reconstitution assays, PA directly activates NADPH oxidase. Phosphatidic acid is also transformed by phospholipase A2to lysophosphatidic acid, an activator of phospholipase C. DAG induces a redistribution of the PKC isoforms beta I, beta II, and zeta from cytosol to the plasma membrane. Human neutrophils express four PKC isoenzymes (alpha, beta I, beta II, zeta). Protein kinase C-beta and -zeta are the most abundant isoforms, whereas PKC-alpha is barely detectable. PMA and DiC8-DAG are synthetic analogs of DAG and induce a similar redistribution and activation of PKC. [9,10]Therefore both probes allow the effects of DAG to be mimicked. However, activation by PMA is maximal and irreversible, whereas the reaction after DiC8-DAG is weaker, reversible, and more comparable to that after FMLP or C5a stimulation. The membrane-bound form of PKC activates NADPH oxidase by intermediate steps. [16,24]Through a single electron transfer, this oxidase transforms molecular oxygen into a superoxide anion. The superoxide anion then dismutates to H2O2inside the phagosomes or in the extracellular space. 
Site of the Nitrous Oxide Interaction
Nitrous oxide decreased the oxidative response. This correlated with a decrease in the number of neutrophils that displayed generation of oxidative products with stimulation with FMLP or C5a (Table 1). The inhibition was concentration dependent (Figure 3).
A reduced number of receptors might lead to a decreased response, as observed. However, no reduction in the number of FMLP receptors was detected, although changes in neutrophil receptor expression have been described after anesthesia. The finding of a comparable [Ca2+]iresponse with and without N2O may suggest an impairment of the response downstream of phospholipase C. Nevertheless, N2O may have an effect on the G proteins or phospholipase C because the intracellularly released Ca2+may not be due solely to inositol-1,4,5-trisphosphate-dependent mechanisms. In this context, an activation of phospholipase C by PA and lysophosphatidic acid or of guanylate cyclase by N2O, resulting in increased intracellular cyclic guanosine monophosphate levels, must be considered potential explanations for the inhibition of the oxidative response. However, the inhibition of the oxidative response, but not of the [Ca2+]iresponse, indicates that N (2) O exerts a postreceptor modulation of signal transduction.
Neither the DiC8-DAG nor the PMA-induced oxidative response was altered in the presence of N2O, which indicates an intact signaling downstream of PKC and an unimpaired function of NADPH oxidase in the presence of N2O. However, neither PMA nor DiC8-DAG require all of the cofactors that DAG requires to activate PKC, and thus a modulation by N2O at the level of PKC activation is not excluded. Saito et al. showed an interference of N2O with PKC. Nevertheless, it should be noted that the tested PKC was isolated from rat neuronal tips, and thus probably belonged to isoforms differing from those found in human neutrophils. Assuming an intact function of PKC, the decreased oxidative response of neutrophils in the presence of N2O might be due to decreased levels of PA or DAG. The former may directly activate NADPH oxidase and has been shown to mediate through its intracellular level the enhanced oxidative response of neutrophils in the presence of proinflammatory cytokines such as tumor necrosis factor-alpha. The enhanced response in the presence of tumor necrosis factor-alpha is associated with a changed ratio between nonresponsive and responsive neutrophils, with more of the latter. Nitrous oxide increases the number of nonresponsive neutrophils, perhaps associated with altered second messenger levels. However, experiments including the direct measurement of the involved second messengers PA, lysophosphatidic acid, DAG, and inositol-1,4,5-trisphosphate (Figure 1) would be needed to substantiate these conclusions. In addition, more defined cellular models such as neutrophils in the absence of other potentially interfering leukocytes, or cells from defined cell lines transfected with the investigated chemotactic receptors are required to characterize the mechanism of interaction at the molecular level.
A scavenging of reactive oxygen derivatives by N2O, a direct interference with H2O2-generating enzymes, or an inhibition of the oxidation or detection of the dye used are also potential causes for an altered H2O2production or detection. Of importance for such unspecific interactions is the stoichiometric relation among oxygen products, N2O, and the dye used. Therefore, the concentrations of 100 nM for FMLP, 200 pM for C5a, and 7.5 micro Meter for DiC8-DAG were chosen to achieve the same level of oxidative response with the different stimuli. These concentrations were also chosen to allow analysis of the heterogeneous response of neutrophils at low degrees of stimulation, and to allow comparison with other authors' data on halogenated volatile anesthetics. [29,30]However, the finding of an unchanged DiC8-DAG- or PMA-induced oxidative response in the presence of N2O excludes these possible causes. This is also supported by the observation of an unchanged spontaneous oxidative burst of monocytes in the presence of N2O (Figure 2(C)). The regulation of this monocyte reaction involves signaling pathways that differ from those mediating neutrophil responses.
Irreversibility of the Nitrous Oxide Effect
A striking result of our study was the irreversibility of the impairment of neutrophil signaling after N2O exposure. That this finding is related to our experimental method can be excluded because, as recently published, the effects of halogenated, volatile anesthetics were completely reversible under the same conditions. This suggests an effect unique to N2O with a sustained change in the properties of one or more of the involved signaling components. Levin and Blanck reported that halothane and isoflurane irreversibly decrease the ability of calmodulin to bind calcium. Nitrous oxide, which was not addressed in their report, might also be able to alter the properties of calmodulin. This would lead to the findings we observed because the proper functioning of calmodulin is necessary for the initialization of the oxidative response. [32,33]A sustained change in protein properties might also occur in relation to the methionine metabolism. Nitrous oxide affects methionine synthase in vitro and in vivo by a one-electron oxidation of cobalamin in the enzyme. [34,35]Further, neutrophils have been shown to have a much higher turnover of methionine than other cell types. As much as 25% of the methionine residues are oxidized in unstimulated neutrophils, whereas oxidized methionine is barely detectable in other cell types. Although the oxidation of methionine residues does not always result in complete protein inactivation, it seems possible that neutrophils exposed to N2O might be especially susceptible to oxidative damage by a sustained alteration of their methionine metabolism.
Neutrophils display a decreased oxidative response to FMLP or C5a stimulation in the presence of N2O. This is characterized by a lower percentage of responsive neutrophils and consequently an overall decrease in H2O2production. Nitrous oxide did not affect the number of receptors, nor did it influence the direct induction of the oxidative response by the PKC activators DiC8-DAG and PMA. This suggests an interference of N2O upstream of NADPH oxidase. Whether the described effects of N2O are relevant for clinical anesthesia is still controversial. However, neutrophils can serve as a good model to study signaling in human cells because the underlying signaling pathways are ubiquitously present in all cells. The further characterization of the N2O effects on cellular signal transduction is of interest to help achieve a better understanding of the molecular mechanisms underlying the pharmacologic action of N2O.