Cytokine upregulation of surface antigens correlates to the priming of the neutrophil oxidative burst response

Authors


Abstract

Background

Neutrophil activation is strongly related to organ dysfunction that occurs during systemic inflammatory responses. The aim of our study was to analyze the oxidative burst response in correlation to the up- and downregulation of N-formyl-L-methionyl-L-leucyl-phenylalanine (fMLP) receptors and the surface antigens CD11b, CD62L, and CD66b as potential surrogate markers of the degree of neutrophil priming for an increased oxidative burst response induced by proinflammatory cytokines.

Methods

Blood was taken from healthy donors. Neutrophils were pretreated with cytokines (interleukin [IL]–1β, IL-6, IL-8, granulocyte-macrophage colony-stimulating factor [GM-CSF], and tumor necrosis factor α [TNFα]; 0.01–10 ng/ml) and stimulated with fMLP (100 nM) in vitro. Functional and phenotypical parameters were quantified flow cytometrically.

Results

The oxidative burst response increased after priming with 0.1 ng/ml TNFα, 1 ng/ml GM-CSF, or 10 ng/ml IL-8. Upregulation of fMLP receptors, CD11b, and CD66b and downregulation of CD62L showed a close correlation to the oxidative burst response. Altered expression of these parameters partly reached significance at lower cytokine concentrations in comparison with the oxidative burst. IL-1β and IL-6 had no effect.

Conclusions

Our results showed that the expression of phenotypical parameters closely correlates with functional parameters in human neutrophils. Thus an up- or downregulation of antigens such as CD11b or CD62L reflects cytokine-induced functional changes. Cytometry Part A 57A:53–62, 2004. © 2003 Wiley-Liss, Inc.

Patients with systemic inflammatory response syndrome (SIRS) or sepsis as defined by the American College of Chest Physicians and the Society of Critical Care Medicine (1) are a challenge for critical care units. Despite many attempts in the past, there is a lack of reliable laboratory parameters to assess the reactions in SIRS and sepsis. In addition, development of multiple organ dysfunction syndrome is hard to predict (1). Although the oxidative burst response (production of reactive oxygen derivatives) of neutrophils is a major host defense mechanism to kill invading microorganisms, a coincidence of excessive neutrophil activation and the development of multiple organ dysfunction syndrome has been shown (2, 3).

The proinflammatory cytokines tumor necrosis factor α (TNFα), interleukin (IL)–1β, IL-6, and IL-8 have been strongly associated with sepsis syndrome and have been used as surrogate markers for the diagnosis of SIRS and sepsis (4). The peak plasma levels of these cytokines detected in patients with inflammatory diseases and with sepsis typically were higher than 3,000 pg/ml for TNFα (5), higher than 1,400 pg/ml for IL-8 (6), and up to 8,000 pg/ml for IL-6 (7). A major problem is the short-lived concentrations of circulating cytokines with half-lives of only few minutes. Many clinical studies have shown a high variability in the plasma levels and a poor correlation between plasma concentrations of proinflammatory cytokines and SIRS or sepsis. Interestingly, however, a persistent elevation of TNFα and IL-6 was found to correlate with the severity of septic shock (8, 9).

Priming of cells describes their ability to adopt an increased functional status. The priming phenomenon was associated with the action of, e.g., cytokines, and priming seems to be closely related to morphologic and phenotypical changes of cells in vitro (10). Therefore, an alternative approach to assess inflammatory reactions during SIRS or sepsis might be the determination of functional or phenotypical parameters in neutrophils instead of quantifying plasma levels of short-lived cytokines.

The oxidative burst response is an important functional parameter characterizing the microbicidal activity of neutrophils (2). A correlation between the oxidative burst response and the expression of receptors for N-formyl-L-methionyl-L-leucyl-phenylalanine (fMLP) in septic patients has been shown in previous studies (11). CD11b (Mac-1) is crucial for adhesion of leukocytes, and it is constitutively expressed in leukocytes but also rapidly upregulated in patients with SIRS (12). CD62L (L-selectin) is a leukocyte-expressed adhesion receptor that plays an important role in directing leukocytes to sites of inflammation (13). In contrast to CD11b and CD62L, CD66b (CGM-6, renamed CEACAM8 in 1999) is used as a degranulation marker. CD66b does not maintain rolling adhesion under flow conditions, but only few data on its detailed role in inflammation exist (14–16).

Although they are technically less complicated to quantify, the correlation of phenotypical markers such as CD11b, CD62L, and CD66b with functional parameters such as the oxidative burst response has not sufficiently been analyzed.

Previous studies often determined single functional or phenotypical parameters in vitro and their kinetics in healthy donors or in patients with SIRS or sepsis (7, 17). A comparative comprehensive analysis of the effect of proinflammatory cytokines on all these parameters under defined in vitro priming conditions has not been performed.

The objective of our in vitro study was to analyze whether cytokine-induced alterations of the oxidative burst response of human neutrophils are reflected by an up- or downregulation of different surface antigens and receptors (CD11b, CD62L, CD66b, and fMLP receptors). Kinetics and dose-response curves upon pretreatment with single proinflammatory cytokines (TNFα, selective TNF receptor subtype agonists, IL-1β, IL-6, IL-8, and granulocyte-macrophage colony-stimulating factor [GM-CSF]; 0.01–10 ng/ml) were determined. We performed a cross-validation of the results to compare the sensitivity of these functional and phenotypical parameters with cytokine priming in an effort to find combinations of parameters that might be more promising than direct analysis of cytokine plasma levels for future studies in patients with SIRS or sepsis.

MATERIALS AND METHODS

Characteristics of Blood Donors

This study was approved by the ethical committee of the University of Regensburg Medical School. After informed consent, venous blood was drawn from healthy donors with no history of infection 2 weeks before the experiments. Blood count and differential leukocyte count were determined before each experiment by using a Technicon H*3 hematology analyzer (Bayer, Tarrytown, NY). All donors (n = 6; mean age, 32 years; age range, 27–38 years) had normal counts.

Leukocyte Preparation and Quantification of H2O2 Production

Leukocyte isolation was carried out by sedimentation of erythrocytes on Histopaque 1.077 (Sigma, Deisenhofen, Germany) separation medium. Heparinized (10 U/ml) whole blood (3 ml) was layered on 3 ml of medium. Erythrocytes aggregated at the interface and settled at room temperature without centrifugation. After 40 min the upper 800 μl of the supernatant leukocyte-rich plasma was withdrawn while avoiding 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.

The supernatant leukocyte-rich plasma was suspended at 1:50 (20 μl leukocyte and 980 μl buffer) in Dulbecco's phosphate-buffered saline (D-PBS, Life Technologies, Eggenstein, Germany) containing calcium and magnesium. The leukocytes were loaded with the fluorogenic substrates dihydrorhodamine 123 and carboxy-seminaphthorhodafluor-1-acetoxymethylester (SNARF1/AM) for 10 min at 37°C (both dyes from Molecular Probes, Eugene, OR). The final concentrations were 1 μM for dihydro-rhodamine 123 and 0.1 μM for SNARF1/AM. After addition of cytokines (0.01–10 ng/ml), samples were incubated for 30 min at 37°C (IL-1β and GM-CSF from Calbiochem-Novabiochem, Bad Soden, Germany; IL-6, IL-8, and TNFα from DPC Biermann, Bad Nauheim, Germany). Mutants with changed binding characteristics for TNF receptors 1 (R1) or 2 (R2) were provided by Dr. H. Loetscher (Pharmaceutical Research, Hoffmann-La Roche, Basel, Switzerland; see Acknowledgments). The R32W-S86T mutant shows a complete loss of its binding activity to the TNF R2 but a retained binding activity to the TNF R1, and the A145R mutant has a substantially impaired (2,500-fold) binding activity to the TNF R1 but a sustained binding to the TNF R2. The difference in molecular weight of the TNF mutants used in our experiments was negligible (18).

Next, fMLP was added to stimulate the H2O2 production by neutrophils (Sigma); after 15 min of incubation at 37°C, the reaction was stopped on ice. Dead cells were counterstained with propidium iodide (Serva, Heidelberg, Germany) at a final concentration of 15 μM. The samples were stored on ice under exclusion of light and measured within 1 h.

For analysis we used a FACScan flow cytometer (Becton Dickinson, San Jose, CA) with argon ion laser excitation at 488 nm, which measured 10,000 cells of each stained sample. Data were acquired and processed with LYSIS-II software. Flow cytometric analysis permits quantification of the fluorescence of each neutrophil. The percentage of non-reacting versus reacting cells can be determined (Fig. 2). After calibration with standard dye beads (Quantum 26, Flow Cytometry Standards Europe, Leiden, The Netherlands), the results obtained for the cellular fluorescence were converted into molecules of equivalent soluble fluorochrome (MESF). These MESF units, used for the absolute quantification of cellular fluorescence, allow interassay and interlaboratory comparison of data. Counterstaining of leukocytes with carboxy-SNARF1/AM was used to improve the discrimination between viable and dead leukocytes. The viable cells were assessed by their enzymatic (SNARF1 esterase) activity, resulting in a cleavage of carboxy-SNARF1/AM to SNARF1, which shows a bright red/orange fluorescence. In addition, dead cells were counterstained with propidium iodide (fluorescence > 600 nm). Neutrophils were identified by their typical side scatter light and forward scatter light patterns and by their esterase activity.

Figure 2.

Flow cytometric analyses of the oxidative burst response in 10,000 leukocytes. The y axis represents the oxidative burst response (measured as rhodamine 123 fluorescence) in arbitrary units, and the x axis indicates the side scatter light pattern of cells: neutrophils, monocytes (mo), and lymphocytes (ly). Left: Unstimulated control. The dashed line separates reacting from non-reacting neutrophils. Right: Reaction after stimulation with 100 nM fMLP. Mean fluorescence of neutrophils is shown in arbitrary units. Fluorescence can be converted into 103 MESF by linear regression analysis by performing calibration analyses with standard dye beads (see Materials and Methods). The percentage of neutrophils above the dashed line in this example is 37%.

Supplementary chemotaxis assays (assay principles based on Migratest, Orpegen Pharma, Heidelberg, Germany) were performed to confirm the activity of IL-8 (results not shown).

Expression of Receptors for fMLP

Analysis of the expression of fMLP receptors was based on the principles of the assay described by Allen and colleagues (19). For staining of fMLP receptors, formyl-Nle-Leu-Phe-Nle-Tyr-Lys (Molecular Probes), a fluorescein-labeled, highly specific analog of fMLP, was used, which fluoresces at 520 nm when excited at 488 nm. Leukocyte-rich plasma was suspended in D-PBS (1:50), and after addition of different concentrations of cytokines (see above), samples were incubated for 30 min at 37°C. Subsequently, the samples were cooled to 4°C to avoid further activation of neutrophils. FLPEP was added in a final concentration of 20 nM, incubated for 20 min at 4°C, and washed twice. Nonspecific binding of fLPEP was determined by addition of excess fMLP (10-4 M) in a parallel set of tubes.

Expression of Neutrophil Surface Antigens

Initial experiments with whole blood samples and samples with leukocyte-rich plasma in parallel showed that the results obtained for the surface marker expression did not show a significant difference. Based on these data, all further experiments were performed with leukocyte-rich plasma. Leukocyte isolation was carried out by a sedimentation of erythrocytes using a separation medium as described above. Leukocyte-rich plasma was diluted 1:50 (20 μl leukocyte and 980 μl buffer) in D-PBS (Life Technologies) containing calcium and magnesium. After addition of cytokines, samples were incubated for 30 min at 37°C. Next samples were gently centrifuged and the volume was reduced to 100 μl. Fluorescence-labeled monoclonal antibodies to CD11b (CD11b/fluorescein isothiocyanate [FITC], mouse immunoglobulin [Ig] G1, clone bear-1; Medac, Hamburg, Germany), CD 62L and CD66b (CD62L/FITC, mouse IgG1, clone DREG 56; CD66b/FITC, mouse IgG1, clone 80 H3; both from Immunotech, Hamburg, Germany) were added, and samples were incubated at 4°C for 15 min. The samples were washed twice and resuspended in 500 μl of D-PBS for flow cytometric analysis.

Statistical and Data Analyses

All experimental data are presented as mean values with the standard error of mean. Lilliefors' test was used to examine normal distribution, and Levene's test was used to check homogeneity of variance. One-way analysis of variance was performed to check differences between the means of cytokine-exposed and control samples assessed in parallel (P < 0.05 and 0.01), followed by a post hoc analysis including Bonferroni′s procedure to consider multiple testing. Effective doses of the cytokines, inducing half-maximum responses (EC50/IC50) were obtained from a logistic fit to the dose-response curve (Origin 5.0, Microcal Software, Northampton, USA).

RESULTS

The oxidative response and the percentage of reacting neutrophils were analyzed after cytokine pretreatment and subsequent fMLP stimulation. Expression of fMLP receptors and surface antigens (CD11b, CD66b, and CD62L) were determined with cytokine pretreatment but without fMLP stimulation.

Oxidative Burst Response

The H2O2 production of neutrophils (oxidative burst response) was quantified by using the indicator dye rhodamine 123. Time kinetics were different for TNFα, GM-CSF, and IL-8, with changes from baseline reaching significance levels within 5, 15, and 30 min depending on the cytokine (Fig. 1A). TNFα induced up to a sevenfold increase in the H2O2 production by neutrophils as compared with control values (Fig. 1B). IL-8 and GM-CSF proved to be less potent than TNFα and significant effects were evoked only by cytokine concentrations of 1.0 and 10.0 ng/ml (Fig. 1B). For the TNFα-induced response, an EC50 of 0.15 ng/ml could be calculated from the dose-response curve. For GM-CSF, the EC50 was 0.40 ng/ml, reflecting the greater potency of TNFα as compared with GM-CSF (Table 5). The TNF R1 agonist induced up to a fourfold increase in H2O2 generation. The EC50 was calculated to be 3.4 ng/ml, which is more than 20 times the EC50 for TNFα (Table 5). IL-1β and IL-6, and the TNF R2 agonist proved to be ineffective (Fig. 1 and Table 1).

Figure 1.

Flow cytometric analyses of neutrophils pretreated with different cytokines and stimulated with 100 nM fMLP. A: Time kinetics of the oxidative burst response (different incubation periods) using cytokine concentrations of 10 ng/ml. The x axis indicates time in minutes. The y axis represents the oxidative burst response given in 103 MESF. B: Oxidative burst response of neutrophils (H2O2 generation measured as rhodamine 123 fluorescence) pretreated with increasing concentrations of different cytokines. The x axis indicates the concentration (ng/ml) of cytokines. C: Time kinetics of the percentage of reacting neutrophils participating in the oxidative burst response. The y axis indicates the percentage of reacting neutrophils. D: Neutrophils participating in the oxidative burst response after pretreatment with increasing concentrations of different cytokines. E: Time kinetics of the expression of fMLP receptors in neutrophils. F: FMLP receptor expression after incubation with increasing concentrations of cytokines. The y axis represents the fluorescence given in 103 MESF.

Table 5. Effective Cytokine Doses Inducing Half-Maximum Neutrophil Responses*
Neutrophil parameterTNFα (ng/mL)TNF R1 agonist (ng/mL)GM-CSF (ng/mL)
  • *

    Cytokine concentrations inducing half-maximum upregulation (EC50) or downregulation (IC50) of functional and phenotypical neutrophil responses were obtained where appropriate from a logistic fit to dose-response curves (Origin 5.0, Microcal Software Inc.). fMLP, N-formyl-L-methionyl-L-leucyl-phenylalanine; GM-CSF, granulocyte-macrophage colony-stimulating factor; TNF, tumor necrosis factor; R, receptor.

Oxidative response/rhodamine fluorescence (EC50)0.153.400.40
Percentage of neutrophils generating H2O2 (EC50)0.020.80 
fMLP receptor expression (EC50)0.020.40.11
CD11b expression (EC50)0.008 0.03
CD62L expression (IC50)0.005 0.03
CD66b expression (EC50)0.09 0.06
Table 1. Oxidative Burst Response and fMLP Receptor Expression of Neutrophils After Pretreatment With TNFα or Selective TNF Receptor Subtype Agonists
CytokineConcentration cytokine (ng/mL)Oxidative burst response (103 MESF).Subset of reacting neutrophils (%)fMLP receptor expression (103 MESF)
  • Fluorescence values for oxidative burst response (measured as rhodamine 123 fluorescence) and fMLP receptor expression of neutrophils are presented in 103 MESF, mean ± standard error of the mean of six independent experiments/donors. Oxidative response and subset of neutrophils were determined with fMLP stimulation, and fMLP receptor expression was determined without fMLP stimulation. fMLP, N-formyl-L-methionyl-L-leucyl-phenylalanine; MESF, molecules of equivalent soluble fluorochrome; TNF, tumor necrosis factor; R, receptor.

  • *

    P < 0.05, control versus cytokine pretreated samples.

  • **

    P < 0.01, control versus cytokine pretreated samples.

TNFαControl34.9 ± 9.837.2 ± 4.9107.6 ± 10.4
 0.0156.3 ± 9.259.3 ± 6.2169.4 ± 11.6**
 0.1134.9 ± 23.6**92.4 ± 0.9**219.0 ± 27.8**
 1.0211.4 ± 28.9**96.4 ± 0.5**233.8 ± 22.5**
 10.0245.9 ± 34.3**96.8 ± 0.5**251.7 ± 17.2**
TNF R1 agonistControl31.7 ± 6.832.7 ± 5.1112.7 ± 22.5
 0.0132.6 ± 8.737.0 ± 4.9140.8 ± 19.3
 0.135.6 ± 10.238.2 ± 4.5154.6 ± 21.4
 1.074.6 ± 18.667.6 ± 6.0**190.5 ± 22.5*
 10.0156.2 ± 36.9*91.8 ± 1.2**232.8 ± 25.7**
TNF R2 agonistControl36.3 ± 8.135.2 ± 5.2116.6 ± 22.6
 0.0136.8 ± 8.742.6 ± 4.8117.6 ± 23.6
 0.133.5 ± 7.538.6 ± 5.4125.0 ± 24.6
 1.040.4 ± 9.645.4 ± 4.3124.8 ± 23.6
 10.041.4 ± 9.141.8 ± 3.6118.6 ± 21.4

TNFα treatment of neutrophils without subsequent fMLP stimulation induced a minor oxidative response as compared with fMLP controls or with results obtained with cytokine priming and fMLP stimulation. The mean oxidative responses obtained with 0.1, 1, and 10 ng/ml TNFα (without fMLP stimulation) were 0.50 ± 0.03, 0.62 ± 0.03, and 0.63 ± 0.03, respectively, when normalized to fMLP controls without cytokine treatment.

Percentage of Neutrophils Generating H2O2

Mean rhodamine fluorescence provides no information about the heterogeneity of neutrophils producing H2O2. Therefore, in addition to the analysis of the cell-associated amount of H2O2 produced by neutrophils, the percentage of neutrophils participating in the oxidative response was determined. In addition to the fluorescence of single cells, the percentage of non-reacting cells versus activated/reacting cells could be analyzed flow cytometrically, as shown in Figure 2.

TNFα, the TNF R1 agonist, IL-8, and GM-CSF mainly exerted their effects within 15 min. Without cytokine pretreatment, 30% to 40% of neutrophils produced H2O2 upon stimulation with 100 nM fMLP. TNFα dose-dependently induced an increase in the percentage of activated neutrophils of up to 97%, the TNF R1 agonist up to 92%, IL-8 and GM-CSF up to 70% (Fig. 1). The EC50 values for TNFα and the TNF R1 agonist were 0.02 and 0.8 ng/ml, respectively (Table 5). IL-1β, IL-6, and the TNF R2 agonist did not induce any significant changes from baseline values (Fig. 1 and Tables 1 and 2).

Table 2. Time Kinetics of the Oxidative Burst Response and the fMLP Receptor Expression After Pretreatment of Neutrophils With TNFα or Selective TNF Receptor Subtype Agonists
Cytokine (10 ng/mL)Time (min)Oxidative burst response (103 MESF)Subset of reacting neutrophils (%)fMLP receptor expression (103 MESF)
  • Fluorescence values for the oxidative burst response (measured as rhodamine 123 fluorescence) and fMLP receptor expression of neutrophils are presented in 103 MESF, mean ± standard error of the mean of six independent experiments/donor. Analogous to data in Table 1, oxidative response and subset of neutrophils were determined with fMLP stimulation, and fMLP receptor expression was determined without fMLP stimulation. fMLP, N-formyl-L-methionyl-L-leucyl-phenylalanine; MESF, molecules of equivalent soluble fluorochrome; TNF, tumor necrosis factor; R, receptor.

  • *

    P < 0.05, control versus cytokine pretreated samples.

  • **

    P < 0.01, control versus cytokine pretreated samples.

TNFα041.7 ± 8.234.7 ± 12.1117.6 ± 16.4
 598.1 ± 12.2**59.4 ± 13.0174.6 ± 20.4**
 15206.4 ± 25.0**93.8 ± 0.4**224.3 ± 22.5**
 30234.8 ± 33.1**96.2 ± 0.6**241.2 ± 24.6**
TNF R1 agonist042.8 ± 9.637.6 ± 11.8119.7 ± 22.5
 590.4 ± 27.159.4 ± 9.8167.3 ± 22.5
 15119.7 ± 32.474.8 ± 12.5224.3 ± 20.8*
 30150.3 ± 32.4*81.0 ± 10.4*239.1 ± 23.6**
TNF R2 agonist046.8 ± 8.636.4 ± 12.4118.6 ± 22.6
 564.7 ± 10.648.6 ± 11.0146.1 ± 17.4
 1561.4 ± 23.942.4 ± 14.0139.8 ± 16.9
 3066.5 ± 28.342.8 ± 13.7144.0 ± 17.0

Expression of fMLP Receptors

TNFα, the TNF R1 agonist, GM-CSF, and IL-8 induced time- and dose-dependent increases in fMLP receptor expression. Significant changes from baseline values were mainly induced within 5 min (Fig. 1E, Table 2). TNFα induced up to a threefold increase of the number of fMLP receptors (Fig. 1F). The TNF R1 agonist and GM-CSF proved to be less potent than TNFα reflected by EC50 values of 0.40 and 0.11 ng/ml versus an EC50 value of 0.02 ng/ml for TNFα (Table 5). IL-8 induced a twofold increase of the receptor expression at 10.0 ng/ml but proved to be ineffective at lower concentrations (Fig. 1F). As shown for the H2O2 production and the percentage of reacting neutrophils, IL-1β, IL-6, and the TNF R2 agonist did not induce any significant effects (Fig. 1 and Tables 1 and 2).

CD11b (Mac-1/CR3)

The expression of the α-subunit of CD11b on the surface of neutrophils was quantified by using a FITC-labeled anti-CD11b antibody. TNFα induced more than a sevenfold and GM-CSF more than a fivefold dose-dependent increase of the CD11b expression (Table 3). Based on a logistic fit of the dose-response curve, EC50 values of 0.008 and 0.03 ng/ml could be obtained for TNFα and GM-CSF, respectively (Table 5). The maximum effects of TNFα and the TNF R1 agonist at 10 ng/ml were comparable (Table 4). IL-8 was less effective than TNFα and GM-CSF in inducing a twofold increase in CD11b at 10 ng/ml.

Table 3. Effects of Proinflammatory Cytokines on the Surface Antigen Expression of Neutrophils
CytokineCytokine concentration (ng/mL)Surface marker expression (103 MESF)
CD11bCD66bCD62L
  • Fluorescence values for neutrophil surface marker expression are presented in 103 MESF, mean ± standard error of the mean of six independent experiments/donor. GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; MESF, molescules of equivalent soluble fluorochrome; TNF, tumor necrosis factor.

  • *

    P < 0.05, control versus cytokine pretreated samples.

  • **

    P < 0.01, control versus cytokine pretreated samples.

TNFαControl13.2 ± 0.213.1 ± 0.224.6 ± 2.5
 0.0156.9 ± 4.6*19.9 ± 1.411.7 ± 2.6
 0.179.9 ± 8.4**46.2 ± 5.3*2.3 ± 0.2**
 1.090.1 ± 7.1**72.9 ± 10.9*0.8 ± 0.1**
 10.092.9 ± 8.9**78.1 ± 10.1**0.5 ± 0.1**
GM-CSFControl14.1 ± 0.712.6 ± 0.330.9 ± 1.3
 0.0123.7 ± 3.214.0 ± 0.729.6 ± 0.2
 0.174.8 ± 11.3**44.7 ± 7.8*10.6 ± 2.2**
 1.090.5 ± 10.8**59.6 ± 11.4*6.2 ± 1.4**
 10.077.8 ± 10.6**48.2 ± 10.5*9.7 ± 1.5**
IL-1βControl16.6 ± 1.017.3 ± 1.127.2 ± 1.1
 0.0116.6 ± 0.318.4 ± 1.126.8 ± 1.2
 0.116.6 ± 0.520.2 ± 1.425.9 ± 1.6
 1.016.5 ± 0.319.9 ± 1.126.9 ± 1.5
 10.017.7 ± 0.218.2 ± 1.223.2 ± 1.9
IL-6Control17.7 ± 1.417.6 ± 0.928.1 ± 1.6
 0.0119.2 ± 2.016.9 ± 1.128.3 ± 1.5
 0.119.2 ± 1.615.5 ± 1.428.5 ± 1.6
 1.027.3 ± 5.818.2 ± 1.124.3 ± 1.1
 10.025.5 ± 2.617.5 ± 1.228.2 ± 0.7
IL-8Control14.5 ± 0.914.0 ± 0.729.8 ± 0.3
 0.0113.0 ± 0.513.1 ± 0.828.0 ± 0.5
 0.112.9 ± 0.313.4 ± 0.727.1 ± 0.8
 1.011.6 ± 0.412.9 ± 0.827.5 ± 0.5
 10.033.9 ± 1.0**18.7 ± 0.324.0 ± 0.6**
Table 4. Effects of TNFα and Selective TNF Receptor Subtype Agonists on the Surface Antigen Expression of Neutrophils
CytokineCytokine concentration (ng/mL)Surface antigen expression (103 MESF)
CD11bCD66bCD62L
  • Fluorescence values for neutrophil surface marker expression are presented in 103 MESF, mean ± standard error of the mean of six independent experiments/donor. MESF, molecules of equivalent soluble fluorochrome; R, receptor; TNF, tumor necrosis factor.

  • *

    P < 0.05, control versus cytokine pretreated samples.

  • **

    P < 0.01, control versus cytokine pretreated samples.

TNFαControl13.2 ± 0.213.1 ± 0.224.6 ± 2.5
 0.0156.9 ± 4.6*19.9 ± 1.411.7 ± 2.6
 0.179.9 ± 8.4**46.2 ± 5.3*2.3 ± 0.2**
 1.090.1 ± 7.1**72.9 ± 10.9*0.8 ± 0.1**
 10.092.9 ± 8.9**78.1 ± 10.1**0.5 ± 0.1**
TNF R1 agonistControl17.0 ± 0.517.2 ± 0.826.4 ± 1.2
 0.0119.5 ± 0.817.0 ± 1.227.4 ± 0.4
 0.126.7 ± 4.319.2 ± 0.629.5 ± 1.7
 1.051.7 ± 6.3**30.2 ± 3.515.4 ± 2.5**
 10.089.7 ± 6.0**49.5 ± 2.0**3.0 ± 0.1**
TNF R2 agonistControl14.4 ± 1.315.5 ± 1.624.6 ± 1.8
 0.0111.8 ± 0.813.1 ± 1.628.0 ± 2.3
 0.110.8 ± 0.612.0 ± 1.429.1 ± 4.6
 1.016.4 ± 1.815.5 ± 1.935.6 ± 3.3
 10.010.9 ± 0.313.8 ± 1.635.0 ± 2.0

CD62L (L-Selectin)

Pretreatment with TNFα led to a downregulation of CD62L, with expression levels approaching zero (EC50 0.005 ng/ml; Tables 3 and 5). Effects induced by GM-CSF were minor compared with those induced by TNFα, with a surface marker expression decreasing to about one-third as compared with levels without cytokine pretreatment (EC50 0.03 ng/ml; Tables 3 and 5). Effects induced by IL-8 were minor compared with those induced by TNFα and GM-CSF, as shown for other parameters.

CD66b (CGM6)

Similar to CD11b and CD62L, an FITC-labeled anti-CD66b antibody was used for the quantification of the surface antigen. TNFα induced a dose-dependent increase (EC50 0.09 ng/ml) up to in excess of fourfold, and GM-CSF induced an increase (EC50 0.06 ng/ml) up to in excess of threefold, in CD66b expression (Tables 3 and 5). The effects of the TNF R1 agonist were minor compared with the results obtained for CD11b and CD62L. IL-1β, IL-6, and the TNF R2 agonist did not induce an altered expression of CD11b, CD62L, and CD66b.

A comparison of the EC50/IC50 values (Table 5) showed that the most sensitive indicators for cytokine effects were CD11b and CD62L. Expression of fMLP receptors and CD66b were less sensitive indicators than were CD11b and CD62L. The least sensitive indicator for neutrophil treatment with TNFα or GM-CSF was the oxidative burst response (rhodamine fluorescence) with an EC50 several-fold higher than that calculated for CD11b (Table 5).

DISCUSSION

As an attempt to find a systematic description of SIRS, Bone suggested a three-step model (20). First, injured tissues produce cytokines to initiate the inflammatory response, leading, e.g., to neutrophil recruitment. Second, small quantities of proinflammatory cytokines are released into the circulation, leading to an activation of endothelial cells and a further recruitment of neutrophils and other cells. The initiated acute-phase response is usually controlled by a simultaneous decrease in proinflammatory mediators and a release of endogenous antagonists. Occasionally, homeostasis is not reestablished and SIRS develops. The resulting degree of neutrophil activation is considered to correlate to the degree of inflammatory tissue destruction (20).

Among the proinflammatory cytokines, TNFα is supposed to be an important mediator of organ injury during sepsis, whereas IL-1 plays a role as an endogenous pyrogen, inducing, e.g., leukocytosis (21). IL-6 also acts as a pyrogen and as a major stimulus to acute-phase protein production, and IL-8 is described as a neutrophil activator (21). Many previous studies have analyzed elevated plasma levels of cytokines in patients, but the results concerning the correlation with clinical outcome were conflicting (22, 23). Some data have suggested that TNFα and IL-6 may be correlated with early events of septic shock, whereas IL-8 might correlate with late events such as death (24). The plasma levels also showed a high variability that presumably is based on the short half-lives of the cytokines and the fact that only excess amounts of the cytokines that are not trapped by target cells and circulating soluble receptors can be detected in the plasma (8, 23).

Alternative approaches for laboratory diagnosis of systemic inflammation could be surrogate markers that reflect the effect of cytokines but are more stable and easy to detect. The oxidative burst response is an important functional response in neutrophils. This response is enhanced upon cytokine priming, and many previous studies have indicated that neutrophil oxidants participate in the pathogenesis of lung failure in patients with, e.g., sepsis (2, 10). In our experiments, we found that the oxidative burst response is enhanced via an increase of the number of activated cells and via an upregulation of fMLP receptors per cell. Previous studies have also shown that the oxidative burst response upon priming with TNFα and stimulation with fMLP is not only due to the regulation of the fMLP receptor expression (25). However, stimulation of fMLP receptors leads not only to an activation of the oxidative burst response but also to a release of, e.g., leukotrienes or myeloperoxidase (26). So the quantification of the oxidative burst response and the expression of fMLP receptors are complementary results for neutrophil activation.

The oxidative response is primarily of major importance at the site of inflammation next to invading bacteria. An early activation in the circulation can be deleterious, leading to a damage of endothelial cells as in acute respiratory distress syndrome (2, 10). We found that the oxidative burst response is less sensitive to cytokine priming than is the regulation of surface antigens, which might reflect a protective mechanism to prevent tissue damage based on an excessive early production of reactive oxygen derivatives.

Among the priming agents tested, TNFα proved to be the most potent proinflammatory cytokine. Activities of TNFα are mediated by two receptors, the TNF R1 (TNF R55, CD120a) and the TNF R2 (TNF R75, CD120b). Neutrophils express comparable amounts of each receptor type (9, 27). Soluble TNFα rapidly binds to the TNF R1 with high affinity and slowly dissociates from the receptor, leading to an efficient activation of the receptor (28, 29). Previous in vitro experiments with selective TNF mutants have shown that the inflammatory effects of TNFα, e.g., the induction of the oxidative response in neutrophils, are mediated mainly by the TNF R1 (30, 31), which was confirmed in our experiments. The role of the TNF R2, the lower affinity receptor for TNF, in cellular responses and its interaction with the TNF R1 are not fully understood because the signaling processes of both receptors are far more complicated than was imagined some years ago (29, 32). The TNF R2 might act as a regulator of the binding of TNF to the TNF R1, although the exact mechanism is unknown (32).

It has been demonstrated in previous studies that cytokine-activated neutrophils undergo functional and phenotypical changes. Yamashiro and associates showed that priming with proinflammatory cytokines such TNFα induce the production and expression of monocyte chemoattractant protein 1 and expression of chemokine genes (33).

Our experiments also confirmed that, apart from changes in functional parameters, changes to the neutrophil phenotype can be seen upon priming with proinflammatory cytokines. TNFα induced a parallel upregulation of CD11b and CD66b and a downregulation of CD62L. A comparison of the EC50 or IC50 values for TNFα and GM-CSF calculated for the induction of different neutrophil reactions shows that the half-maximum up- or downregulation of the surface antigens is induced by several-fold lower cytokine concentrations than the half-maximum oxidative response. A parallel upregulation of fMLP receptors and CD11b as described previously (34) was confirmed in our experiments.

The most sensitive indicators for cytokine pretreatment proved to be CD11b and CD62L. These neutrophil surface antigens are important for neutrophil recruitment and transmigration through the endothelial layer into tissues in early stages of inflammatory processes. Compared with CD11b and CD62L, the upregulation of CD66b was less sensitive to cytokine pretreatment, which might reflect the role of this antigen as a regulator of the expression of other antigens (e.g., integrins) and its minor role as a direct activator of anti-inflammatory processes (16, 35). There are only a few publications describing the effect of cytokines on the expression of CD66b. De Haas et al. found elevated expressions of CD66b and CD11b in healthy volunteers after injection of GM-CSF, with a peak CD66b expression 30 min after the injection (36). Detailed data describing the effect of TNFα, GM-CSF or other proinflammatory cytokines on the expression of CD66b have not been published. Many data have indicated the pathophysiologic relevance of the adhesion of neutrophils via expression of CD11b. In patients with posttraumatic acute respiratory distress syndrome, pulmonary function, plasma levels of TNFα, and expression of CD11b on neutrophils were closely related (12, 37). CD11b expression was enhanced on circulating neutrophils in patients with systemic inflammatory syndromes, and upregulation of CD11b correlated with organ failure (e.g., of the liver) (34, 38). In contrast, a diminished CD11b activation in response to TNFα in patients with SIRS has been described (7). In our experiments, upregulation of CD11b and downregulation of CD62L were more sensitive to low cytokine concentrations than were the other parameters. Therefore, distinct changes in the expression of CD11b and CD62L even upon priming with small doses of TNFα and GM-CSF (0.01 or 0.1 ng/ml) underline the high sensitivity of these parameters for monitoring cytokine effects.

Summarizing our results, we found that the phenotypical parameters CD11b and CD62L proved to be sensitive surrogate markers to monitor the effects of proinflammatory cytokines, especially TNFα and GM-CSF, in vitro. The major objective for the future is to evaluate the phenotype-function relationship in septic patients to further characterize mechanisms of cytokine priming in sepsis.

Acknowledgements

The selective TNF receptor agonists were provided by Dr. H. Loetscher, Pharmaceutical Research, Hoffmann-La Roche, Basel, Switzerland.

Ancillary