Effect of adhesion on inducible nitric oxide synthase (iNOS) production in purified human neutrophils

Abstract

The production of nitric oxide (NO) within neutrophils is an important element of the innate immune response. We have previously shown that cytokines (IL-1α, tumour necrosis factor-alpha and interferon-gamma) induce human neutrophils in buffy coat preparations to produce iNOS. In order to define better the exact requirements for iNOS production within human neutrophils, we have studied the conditions needed for the production of iNOS in purified neutrophils. In contrast to buffy coat preparations, purified neutrophils in suspension did not produce an increase in iNOS following addition of cytokines. However, when purified neutrophils were allowed to adhere to glass surfaces either uncoated or coated with fetal calf serum (FCS), plasma, fibronectin or laminin, there was an increase in the percentage of iNOS-positive cells. The addition of cytokines during adhesion of these cells increased this proportion further. This was most marked for glass alone and FCS-coated glass on which the proportion of iNOS-positive cells increased to 22·7% and 35·5%, respectively, a significant increase compared with cytokine-treated neutrophils in suspension. Neither transmigration through activated endothelial monolayers nor the addition of soluble intercellular adhesion molecule-1 to purified neutrophil suspensions increased the percentage of iNOS-positive cells following cytokine stimulation. Adhesion of neutrophils to surfaces coated with IgG or complement also failed to increase cytokine-induced iNOS production. We conclude that iNOS production in human neutrophils requires not only cytokine stimulation, but also additional stimuli from adhesion to a surface.

Introduction

The production of nitric oxide (NO) within phagocytes is an important component of the innate immune response to infection. NO has microbicidal activity, which is considerably enhanced by reaction with superoxide anions to yield the highly reactive and microbicidal anion, peroxynitrite [1,2]. Within phagocytes, high output of NO is produced from the enzyme-inducible NO synthase (iNOS), produced following stimulation by agents such as proinflammatory cytokines and lipopolysaccharide (LPS) [3,4].

The role of NO in rodent phagocytes has been well demonstrated by a number of experimental approaches (reviewed in [5,6]). Targeted deletion of the iNOS gene renders mice susceptible to a number of pathogens including Mycobacterium tuberculosis [7], Leishmania major [8] and Listeria monocytogenes [9] (reviewed in [10,11]). Mice deficient in both phagocyte oxidase and iNOS develop spontaneous bacterial abscesses, not seen in animals deficient in either microbicidal system alone [12]. Rodent neutrophils produce large amounts of iNOS following cytokine and LPS stimulation [13].

The role of iNOS in human neutrophils is less clear, with some groups failing to detect any NO release by human neutrophils despite appropriate stimulation [14–16]. However, increased iNOS mRNA expression has been recorded in neutrophils from patients with sepsis, which correlated with an increased nitrite/nitrate concentration in blood [17]. iNOS activity has also been demonstrated in neutrophils isolated from patients with bacterial infections [18,19]. We have previously shown that the addition of interferon-gamma (IFN-γ), tumour necrosis factor-alpha (TNF-α) and IL-1α to human buffy coat preparations induces 20% of neutrophils to produce iNOS mRNA and protein [20]. iNOS co-localized with myeloperoxidase in primary granules and activity was demonstrated indirectly by detection of nitrotyrosine around phagocytosed bacteria [20].

In order to define better the exact requirements for iNOS production within human neutrophils, we have studied the conditions needed for the production of iNOS in purified neutrophils. We find that cytokines alone are not sufficient for the production of iNOS in purified human neutrophils, but that adhesion to a surface in combination with cytokine treatment allows efficient iNOS induction.

Materials and methods

Preparation and culture of neutrophils

Neutrophils were isolated from healthy donors and purified using hypotonic Percoll discontinuous density gradients [21]. The experiments described in the study were all performed using neutrophils from at least two different donors. Microscopic evaluation was used to confirm cell preparations comprised at least 97% neutrophils. Cells were cultured in RPMI 1640 medium (Life Technologies, Paisley, UK) containing 2 mm glutamine and 10% fetal calf serum (FCS) (LabTech Int, Ringmer, UK). Neutrophils (approximately 10⁶/ml) were incubated for 3 h at 37°C in a humidified incubator containing 5% CO₂. Human cytokines (all from Peprotech, London, UK) were added where stated at the following concentrations: 0·5 ng/ml IL-1α, 10 ng/ml TNF-α and 500 U/ml IFN-γ. When neutrophils were allowed to adhere to a surface, the cytokines were added at the same time and were present throughout the incubation. For adherence studies 13-mm diameter glass coverslips (Merck, Leicester, UK) were coated in 200 μl FCS (not heat-inactivated), 200 μl fibronectin (10 μg/ml; Boehringer Mannheim, Lewes, UK), 200 μl human platelet-poor plasma (PPP), or 200 μl laminin (10 μg/ml; Becton Dickinson, Cowley, UK), and incubated for 2 h at 37°C, before washing three times with normal saline. After incubation, coverslips were washed very gently three times with PBS (0·01 m sodium phosphate buffer, pH 7·3, 0·15 m NaCl), then allowed to dry before immunostaining. When determining the optimum time for iNOS induction, apoptotic neutrophils were assayed by nuclear morphological criteria.

Immunostaining

iNOS protein was detected by immunocytochemistry using a murine iNOS MoAb (N39120, clone 54; Transduction Labs, Lexington, KY). This antibody recognizes a single 135-kD protein in cytokine-treated human renal epithelial cells (data not shown), the expected molecular weight of iNOS [22]. Immunostaining was carried out as described [20]. Briefly, cells were fixed in 1% paraformaldehyde for 30 min and endogenous peroxide was blocked by immersion in 0·03% hydrogen peroxide in methanol for 15 min. Non-specific binding was blocked by incubation in 10% horse serum (Serotec, Oxford, UK) in PBS for 30 min at room temperature or 16 h at 4°C. iNOS antibody (1·25 μg/ml) was added in 10% horse serum and incubated for 4 h at room temperature or 16 h at 4°C. Cells were washed in PBS and incubated with 10 μg/ml biotinylated horse antiserum to mouse IgG (Vector Labs, Peterborough, UK). Cells were stained by incubating for 30 min in freshly prepared avidin-biotinylated-peroxidase complex (ABC) reagent (Vectastain; Vector Labs). Peroxidase activity was revealed using 0·03% hydrogen peroxide and 0·35 mm diaminobenzidine treatment, which gives brown staining. Staining was visualized on a Nikon E600 light microscope. Isotype-matched immunoglobulin controls were used in each experiment to check antibody specificity.

Transmigration assay

Transwell inserts (6·5 mm diameter) of 3 μm pore size (Corning Costar, High Wycombe, UK) were precoated with 50 μg/ml fibronectin for 1 h. Human umbilical vein endothelial cells (HUVEC; BioWhittaker, Wokingham, UK) at passage 3–4 were seeded at a density of 10⁵ cells/well in Endothelial Growth Media-2 (EGM-2; BioWhittaker). This medium contains defined growth factors and serum supplements (2% FCS) to support the growth of these cells. HUVEC were cultured for 6 days, changing the growth medium daily. To ensure that a confluent monolayer with tight intercellular junctions had formed, only monolayers with a transendothelial electrical resistance of at least 20–25 Ω.cm² were used in assays [23]. Resistances were measured using an STX-2 electrode (World Precision Instruments, Stevenage, UK). HUVEC were stimulated for 16–24 h with cytokines (0·5 ng/ml IL-1α, 500 U/ml IFN-γ, 10 ng/ml TNF-α) in EGM-2. Transmigration experiments were performed in RPMI 1640 medium containing 2 mm glutamine, 10% FCS and human cytokines at the concentrations described above for iNOS stimulation. Neutrophils (1 × 10⁵) were added above the endothelial layer with either 1 μm N-formyl-methionyl-leucyl-phenylalanine (fMLP) or 25 ng/ml IL-8 (Peprotech) in the lower compartment and incubation continued at 37°C. Transmigrated neutrophils were collected over a period of 3–4 h in low-attachment wells (Corning Costar) to limit attachment to well walls. These were pooled with transmigrated neutrophils that had adhered to the underside of the Transwell membrane, which were released by incubation in 0·025% EDTA solution for 5 min at 4°C. Cells from three to four wells were pooled and washed once in PBS before making a cell smear and immunostaining as described. Cells adherent to the surface of endothelial cells were released by incubating the top surface of the endothelial cells in EDTA and immunostaining for iNOS as described above. Fibronectin-coated membranes without an endothelial layer were used as controls.

Coating surfaces with iC3Bi and IgG

Glass surfaces (either coverslips or eight-well glass TiterTeks (Life Technologies) were coated with iC3Bi according to [24]. Surfaces were treated with 0·1 mg/ml poly l-lysine for 30 min at room temperature and then 2·5% gluteraldehyde was added for 15 min, washing with PBS three times between treatments. After a further washing step, a mixture of 50 μg/ml protein A (Sigma, Poole, UK) and 50 μg ml human IgM (Sigma) was added to surfaces and incubated for 4 h at room temperature. Again the surfaces were washed with PBS and then 2 mg/ml casein was added and left on the surfaces for 16 h at 4°C. After washing in PBS, 10% heat-inactivated FCS was added to the glass and incubated for 2 h at room temperature. The surfaces were washed in PBS and then incubated for 90 min at room temperature in veronal buffer (ICN Biomedicals, Basingstoke, UK) with 5% fresh human serum, 1 mm MgCl₂, 0·15 mm CaCl₂ and 0·1% gelatin. Surfaces were washed once with PBS before adding neutrophils. To coat surfaces with IgG molecules the method of Zhou et al. [25] was used. Coverslips or TiterTeks were coated with 1 mg/ml bovine serum albumin (BSA) solution for 30 min at 37°C. They were washed with PBS and then the surfaces were coated with 0·3 mg/ml rabbit anti-BSA IgG (ICN Biomedicals) for 1 h at room temperature. The glass was washed once in PBS and used for assay. Efficient coating of these surfaces was checked by ELISA. To test for iC3Bi coating, 3·5 μg/ml donkey antibody to human C3 (ICN Biomedicals) were added to the coated coverslips and incubated for 1 h at room temperature. After washing in PBS with 0·1% Tween, a 1:2000 dilution of peroxidase-conjugated horse anti-donkey antibody was added (Jackson ImmunoResearch Labs, West Grove, PA) and incubated for 1 h before washing and testing for peroxidase formation. To test for Fc-coated surfaces a 1:2000 dilution of peroxidase-conjugated goat anti-rabbit antibody peroxidase (Jackson ImmunoResearch Labs) was added to surfaces and incubated for 1 h before washing and measuring peroxidase production.

Soluble intercellular adhesion molecule-1 treatment

Soluble intercellular adhesion molecule-1 (ICAM-1; 75 ng/ml; R&D systems, Abingdon, UK) was added to purified neutrophils in suspension and incubated for 3 h at 37°C.

Statistical analysis

Differences between treatments were assessed by non-parametric tests as variances of the different groups differed markedly. A significant difference between treatments was assessed by a Kruskal–Wallis test, and the contribution of individual treatments to any significant difference found was assessed by individual Mann–Whitney tests, using the Bonferroni correction for multiple comparisons. P < 0·05 was considered significant.

Results

Cytokines do not induce iNOS protein production in purified human neutrophils incubated in suspension

Initial results indicated the optimum incubation time for detection of iNOS production was 3 h; extended incubations led to increased numbers of apoptosed neutrophils. Figure 1 shows iNOS production in purified neutrophil suspensions with and without cytokine stimulation. The percentage of iNOS-positive cells in unstimulated suspensions of purified neutrophils was 6·39 ± 1·75% (mean ± s.e.m., n = 7, range 0·9–12%) (Figs 1 and 2). In contrast to our previous results with human buffy coat preparations, addition of IL-1α, TNF-α and IFN-γ did not induce iNOS production and numbers of positive cells remained low at 4·61 ± 1·57% (n = 7, range 0·4 – 10%) (Figs 1 and 2).

Fig. 1.

iNOS protein in human neutrophils incubated in suspension for 3 h, (A) without cytokine stimulation, (B) with the addition of IFN-γ, IL-1α and tumour necrosis factor-alpha during incubation. Mag. × 1000.

Fig. 2.

Effect of adhesion to the indicated surfaces on iNOS production in purified neutrophils. □, Cells in medium alone; ▪, results from cells incubated in the presence of IFN-γ, IL-1α and tumour necrosis factor-alpha. Each bar represents the mean of four separate experiments; error bars are ± s.e.m. At least 200 neutrophils were counted to obtain the percentage of cells that expressed iNOS. *Significant difference from neutrophils in suspension, with cytokines (Mann–Whitney test with Bonferroni correction, P < 0·05).

Effects of adhesion to a surface on iNOS induction

Previously we have found cytokines stimulated 20% of the neutrophils in buffy coat cell suspensions to make iNOS protein [20]. Addition of cell-free supernatants from cytokine-stimulated buffy coat suspensions to purified neutrophils in suspension did not induce iNOS protein production over that seen with no additions (data not shown). Thus, the presence of additional soluble factors within the cytokine-stimulated buffy coat suspension was not responsible for the efficient induction of iNOS in these cells. Buffy coat suspensions undergo marked homotypic aggregation upon cytokine stimulation, which we did not observe in our purified neutrophil suspensions. Given this observation, we investigated the effect of adherence on iNOS production in purified neutrophils. We allowed neutrophils to adhere to glass coverslips and then directly immunostained for iNOS after 3 h incubation. The effect of adhesion to glass and glass coated with FCS, human plasma, fibronectin and laminin was studied. In addition, we tested the effect of the addition of the cytokines IL-1α, TNF-α and IFN-γ added during this 3-h adhesion period.

Results showed a general trend of increased iNOS production upon adhesion, compared with neutrophils maintained in suspension (Fig. 2). Addition of cytokines during adhesion tended to increase the proportion of iNOS-positive cells further. Since the variance of the groups differed markedly, we used a non-parametric analysis of the significance of the differences between the groups. This showed that there were significant differences between the cytokine-treated groups (P = 0·0187, Kruskal–Wallis test). To determine which treatments accounted for this difference between the cytokine-treated neutrophils, we performed Mann–Whitney tests of significance between the results for adherence to each of the surfaces indicated and cells in suspension. This showed that cells treated with cytokines during adhesion to glass alone and to FCS-coated glass (Fig. 3) had a significantly higher percentage of iNOS-positive cells compared with cytokine-treated cells in suspension (P < 0·05, Mann–Whitney test with Bonferroni correction for multiple comparisons). These differences were not due simply to differences in the numbers of cells that adhered to the glass coated with the different substrates. The percentages of cells that adhered to each surface from the total added was relatively similar for each surface, as shown in Table 1.

Fig. 3.

iNOS protein in human neutrophils adhered to fetal calf serum-coated coverslips incubated for 3 h, (A) without cytokine stimulation, (B) with the addition of IFN-γ, IL-1α and tumour necrosis factor-alpha during incubation. Mag. × 1000.

Table 1.

Percentages of total numbers of neutrophils added adhering to the indicted surfaces

Treatment	Percentage of cells adhering (mean ± s.e.m.)
Glass	64 ± 4·1
Glass + cytokines	70 ± 8·1
FCS-coated glass	58 ± 8·7
FCS-coated glass + cytokines	60 ± 7·1
PPP-coated glass	86 ± 2·9
PPP-coated glass + cytokines	84 ± 1·7
Fibronectin-coated glass	66 ± 5·4
Fibronectin-coated glass + cytokines	79 ± 4·6
Laminin-coated glass	ND
Laminin-coated glass + cytokines	ND

Results are shown as the means and standard errors calculated from at least three experiments.

FCS, Fetal calf serum; PPP, platelet-poor plasma; ND, Not determined.

Endothelial transmigration does not induce iNOS protein production

We investigated the effect of transmigration through endothelial monolayers on iNOS protein production. Confluent endothelial monolayers were stimulated with IL-1α, IFN-γ and TNF-α for 16 h. Neutrophils were added above the stimulated endothelial monolayer and fMLP or IL-8 was used as chemoattractant in the lower compartment. Cytokines were added to both compartments. Neutrophils that had transmigrated through an endothelial monolayer in the presence of cytokines did not have a significantly increased proportion of iNOS-positive cells compared with cytokine-treated suspensions of the same neutrophil preparations (Fig. 4). In addition, those cells that had adhered to the endothelial monolayer showed no increase in the percentage of iNOS-positive cells (Fig. 4). Neither IL-8 nor fMLP had any effect on the production of iNOS in cytokine-treated neutrophil suspensions.

Fig. 4.

Effect of transmigration through a confluent endothelial cell layer on iNOS production in neutrophils. The treatments are as follows: S, cells in suspension; S + C, cells in suspension + cytokines; S + C + IL-8, cells in suspension + cytokines + IL-8; S + C + N-formyl-methionyl-leucyl-phenylalanine (fMLP), cells in suspension + cytokines + fMLP; T (IL-8), transmigrated neutrophils using IL-8; T, mo (IL-8), transmigrated neutrophils across membrane alone using IL-8; T (fMLP), transmigrated neutrophils using fMLP; T, mo (fMLP), transmigrated neutrophils across membrane alone using fMLP; A (fMLP), neutrophils adherent to endothelial surface with fMLP as chemoattractant. Means of two to seven separate experiments using three different donors are shown; error bars indicate ± s.e.m. Percentages were calculated by counting at least 200 neutrophils.

Adherence to surfaces coated with IgG or complement does not induce iNOS production

On adherence to coated glass surfaces in the presence of cytokines, purified human neutrophils show an increase in the proportion of iNOS-positive cells. To investigate the molecular nature of this interaction, we tested the effect of neutrophil binding to ligands important in neutrophil adhesion and activation. We investigated the effect of addition of sICAM-1 and adhesion to complement and IgG-coated surfaces, either with or without the addition of cytokines. None of these treatments induced iNOS production in purified human neutrophils over that found in neutrophils in suspension (Table 2).

Table 2.

Effect of adhesion to surfaces coated with IgG or iC3Bi and addition of soluble intercellular adhesion molecule-1 (sICAM-1) on iNOS production in purified human neutrophils

Treatment	Percentage iNOS-positive neutrophils (mean ± s.e.m.)
Suspension without cytokines	10·3 ± 3·6
Suspension with cytokines	6·5 ± 1·9
Suspension with sICAM-1	8 ± 2·5
Suspension with sICAM-1 and cytokines	7·9 ± 5·6
Adhesion to iC3Bi without cytokines	6·6 ± 2·9
Adhesion to iCBi with cytokines	8·4 ± 2·8
Adhesion to IgG without cytokines	3·8 ± 2·8
Adhesion to IgG with cytokines	5·8 ± 1·5

Purified human neutrophils were incubated for 3 h alone or in the presence of 0·5 ng/ml IL-1α, 10 ng/ml tumour necrosis factor-alpha and 500 U/ml IFN-γ.

Results are shown as the means and standard errors calculated from at least three experiments.

Discussion

In contrast to our previous results with buffy coat suspensions [20], the present results show that purified neutrophils in suspension are not induced to produce iNOS with the addition of cytokines (Figs 1 and 2). Addition of cell-free supernatants from cytokine-stimulated buffy coat suspensions to purified neutrophils in suspension did not induce iNOS protein production. This suggests soluble factors released by other cells are not responsible for stimulating iNOS production in human buffy coat suspensions. One noticeable difference between our suspensions of purified neutrophils and buffy coat cells is that buffy coat cells show much greater homotypic cell aggregation following cytokine stimulation, a process mediated by l-selectin and β₂-integrins [26–28]. Aggregation and adherence of neutrophils can result in them becoming more responsive to soluble ligands, for example TNF-α and -β [29] and LPS [30]. In addition, adhesion causes intracellular changes such as reorganization of the cytoskeleton structure and Ca²⁺ oscillations [31] and can also alter the way a neutrophil responds to different stimuli.

We found that neutrophils adherent to glass and glass coated with FCS, PPP, fibronectin, or laminin showed a trend to increased iNOS production even in the absence of cytokine stimulation (Fig. 2). The addition of cytokines during adhesion of neutrophils to a variety of surfaces produced further increases in the numbers of iNOS-positive cells. We observed high variability of the percentages of iNOS-positive neutrophils, which could partly, but not solely, be attributed to differences between neutrophils from different donors. Overall comparison of the data shown in Fig. 2 using the Kruskal–Wallis test showed that there was a significant difference between the different cell treatments. Individual comparison was made between cells treated with cytokines in suspension and those allowed to adhere to five different surfaces in the presence of cytokines. After correction using the Bonferroni method for multiple comparisons, we found the differences observed only reached statistical significance for the cells adhered to glass and FCS-coated glass (Fig. 2). The results show that cytokines alone are not sufficient to induce iNOS production within purified neutrophils in suspension. However, the additional stimulus of adherence to a glass or FCS-coated glass surface in the presence of cytokines allows efficient iNOS induction.

What processes are responsible for mediating the increased iNOS induction after adherence to these surfaces in the presence of cytokines? Serum is a complex biological fluid with many components, including extracellular matrix proteins such as fibronectin and laminin. However, these proteins alone failed to provide the additional stimulus necessary for iNOS induction (Fig. 2). We also speculated that glass and FCS-coated glass might provide a surface to which LPS could bind, providing an additional stimulus for iNOS induction. LPS binding to serum-coated surfaces has been demonstrated to produce neutrophil activation [32]. However, when we added LPS directly to glass surfaces prior to the addition of neutrophils, we saw no augmentation of iNOS induction with cytokines (data not shown).

A more biologically relevant surface to which neutrophils attach is the endothelium, prior to transmigration to sites of tissue injury. This process involves many different receptor–ligand interactions, including integrins and selectins [33–35]. Therefore, we studied the effect of transmigration on iNOS induction. Surprisingly, we found there was no increase in numbers of iNOS-positive neutrophils after transmigration through a cytokine-treated endothelial layer, even with the additional stimulus of cytokines (Fig. 4). Neutrophils that adhered to this endothelial layer similarly did not show an increase in numbers of iNOS-positive cells. Indeed, the percentage of iNOS-positive cells was lower than in those cells treated in suspension, although the absolute reduction in numbers was small compared with the increases seen following adhesion (Fig. 2).

To explore further the cell surface receptor involved in induction of iNOS production upon adhesion, we investigated the effect of adding sICAM-1 to neutrophils in suspension. ICAM-1 binds to neutrophils via the β₂-integrins CD11a/CD18 and CD11b/CD18 [36] and is expressed on endothelial cells and other parenchymal cells [37]. Binding of CD11b to ICAM-1 promotes PMN adhesion, migration and respiratory burst [38] and binding to CD18 causes elastase release [39]. We found that addition of sICAM-1 did not induce iNOS production in neutrophils in suspension, alone or in combination with cytokines (Table 2). We also investigated the effect of adherence to IgG or complement-coated surfaces. Neutrophils bind the Fc domains of IgG via FcγRII (CD32) and FcγRIII (CD16) receptors [40] and complement C3Bi via CD11b/CD18 [41]. These receptors are involved in many neutrophil functions, including phagocytosis, endothelial transmigration and activation. However, we found that the percentage of iNOS-positive neutrophils did not increase significantly following adherence of cells to these surfaces, either with or without the addition of cytokines (Table 2). Despite stimulating a range of neutrophil receptors involved in adhesion and tissue migration, we have been unable to identify a neutrophil receptor that might mediate the increased cytokine-induced production of iNOS on adhesion to a surface.

Neutrophils isolated from the peripheral blood of healthy donors are not primed for optimal iNOS production. Our results show that adhesion to uncoated or FCS-coated glass surfaces together with cytokine stimulation induce purified human neutrophils to synthesize iNOS. The induction of iNOS upon adhesion does not appear to involve β₂-integrins, selectins or Fcγ receptor; the additional signals required remain unknown. We speculate that the requirement for adhesion in addition to cytokine treatment for efficient iNOS induction will lessen the possibility of iNOS production in neutrophils in the circulation. This may target production of neutrophil iNOS to cells aggregated or adhered in the vicinity of an inflammatory focus.