Eosinophils adhere to airway cholinergic nerves and influence nerve cell function by releasing granule proteins onto inhibitory neuronal M2 muscarinic receptors. This study investigated the mechanism of eosinophil degranulation by cholinergic nerves. Eosinophils were cocultured with IMR32 cholinergic nerve cells, and eosinophil peroxidase (EPO) or leukotriene C4 (LTC4) release was measured. Coculture of eosinophils with nerves significantly increased EPO and LTC4 release compared with eosinophils alone. IMR32 cells, like parasympathetic nerves, express the adhesion molecules vascular cell adhesion molecule-1 and intercellular adhesion molecule-1 (ICAM-1). Inhibition of these adhesion molecules alone or in combination significantly inhibited eosinophil degranulation. IMR32 cells also significantly augmented the eosinophil degranulation produced by formyl-Met-Leu-Phe. Eosinophil adhesion to IMR32 cells resulted in an ICAM-1-mediated production of reactive oxygen species via a neuronal NADPH oxidase, inhibition of which significantly inhibited eosinophil degranulation. Additionally, eosinophil adhesion increased the release of ACh from IMR32 cells. These neuroinflammatory cell interactions may be relevant in a variety of inflammatory and neurological conditions.
- reactive oxygen species
eosinophils play an important role in allergic disorders such as asthma, atopic dermatitis, and allergic gastrointestinal diseases as well as in the elimination of parasitic infections and as part of the immune response to some cancers (10). Eosinophils release a wide variety of mediators, of which the granular proteins eosinophil peroxidase (EPO), major basic protein (MBP), and eosinophil-derived neurotoxin appear to play a pivotal role in these conditions (3).
In the lungs, stimulation of parasympathetic nerves in the vagus leads to bronchoconstriction via ACh binding to M3 muscarinic receptors on the airway smooth muscle (24). Control over the release of ACh is mediated by neuronal muscarinic M2receptors in an autocrine manner (9). In antigen-sensitized guinea pigs after antigen challenge and in some asthmatic patients, there is loss of function of pulmonary neuronal M2 muscarinic receptors that leads to increased vagally mediated bronchoconstriction (4, 8). In vitro ligand-binding studies have shown that eosinophil MBP and EPO are antagonists at M2 muscarinic receptors (15). Previously, we have demonstrated that, in vivo, there is a specific localization of eosinophils and extracellular MBP to airway cholinergic nerves (2, 5). Furthermore, inhibition of eosinophil MBP preserves M2 receptor function in antigen-challenged guinea pigs (6). Thus these data suggest that eosinophils localize to cholinergic nerves and release MBP, which in turn is associated with enhanced ACh release from the nerves.
Because eosinophil granular proteins do not disperse after their release and therefore act primarily in the vicinity of the eosinophil, it is likely that the signal to degranulate is similarly localized. Previous studies have shown that the release of eosinophil granular proteins is enhanced by adhesion through specific adhesion molecules (21). We have shown that primary cultures of airway parasympathetic cholinergic nerves and the human cholinergic neural cell line IMR32 express the adhesion molecules intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), to which eosinophils adhere (27). In this study, we investigated the hypothesis that this adhesion of eosinophils to nerves leads to eosinophil degranulation and thus leads to increased ACh release from the nerves.
MATERIALS AND METHODS
DMEM, Trowell's T8 medium, and FCS were purchased from GIBCO-BRL Life Technologies (Paisley, UK), and tissue culture plasticware was from Helena Biosciences (Sunderland, UK). The IMR32 cell line was obtained from ECACC (Salisbury, UK). Heparin was from Leo Laboratories (Princes Risborough, UK), and Percoll and [methyl-3H]choline were purchased from Amersham Pharmacia Biotech (St. Albans, UK). CD16 immunomagnetic beads and VS+ VarioMacs columns were from Miltenyi Biotech (Bisley, UK). DiffQuick Fix was from Dade (Munich, Germany). The leukotriene C4 (LTC4) enzyme immunoassay (EIA) kit was from Cayman Chemicals (Ann Arbor, MI). Homovanillic acid (HVA), paraformaldehyde, dihydrorhodamine 123 (DHR), penicillin/streptomycin solution, gentamicin, sodium butyrate, PBS, diphenyleneiodonium chloride (DPI), apocynin, and all common buffer constituents were obtained from Sigma (Poole, UK). N-formyl-Met-Leu-Phe (fMLP) peptide was from Peninsula Laboratories Europe (St. Helens, UK). The peroxidase substrate kit (Vector) was from Vector Laboratories (Peterborough, UK), and the peroxidase-conjugated immunoglobulin standard was purchased from Dako (High Wycombe, UK). Mouse anti-human CD11/18 (MEM 48, isotype IgG1) monoclonal antibody (MAb) was from Accurate Scientific Chemicals, mouse anti-human CD18 MAb (685A5, isotype IgG2A) was from Cymbus Biotechnology (Chandlers Ford, UK), mouse anti-human VCAM-1 MAb (B-K9; isotype IgG1) was from Lab Vision, goat anti-human p47phox (N-19) polyclonal antibody was from Autogen Bioclear (Calne, UK), and the isotype controls mouse IgG1and mouse IgG2A were from Caltag Laboratories (Silverstone, UK). Mouse anti-human ICAM-1 MAb (RR1/1.1.1) was a gift from Dr Robert Rothlein (Boehringer-Ingelheim Pharmaceuticals), and ZD-7349 was a gift from Dr. Duncan Haworth (Astra Zeneca Pharmaceuticals, Macclesfield, UK).
Eosinophils were prepared from the blood of healthy human volunteers by a negative immunomagnetic selection technique. After phlebotomy, 15 ml of blood were added to 25 ml of PBS containing 100 units of heparin and then were layered on a 1.090 g/ml Percoll solution. After centrifugation at 400 g for 20 min, the upper layer and mononuclear cells were discarded, and the pellet containing granulocytes and red blood cells was exposed to hypotonic lysis by treatment with 18 ml of H2O for 30 s followed by the addition of 2 ml PIPES-buffered salt solution (250 mM PIPES, 1.1 M NaCl, 50 mM KCl, and 420 mM NaOH; pH 7.47) to restore tonicity. After centrifugation at 400 g for 6 min, the pellets were pooled, resuspended in a one-tenth dilution of the above PIPES-buffered salt solution containing 55.5 mM glucose and 3% (wt/vol) human serum albumin, and washed by centrifugation, and the hypotonic lysis procedure was repeated. The remaining granulocytes were then resuspended in PBS plus 2 mM EDTA and 0.5% BSA (MACS buffer) with 1 μl of CD16 immunomagnetic beads/106 cells and incubated for 30 min at 4°C. The cells were then washed, resuspended in MACS buffer, and passed through a magnetic separation column (Miltenyi Biotech). The immunomagnetically labeled CD16-positive neutrophils were retained in the column, and the eluted cells were collected and centrifuged. The pellet was resuspended in differentiation medium (DMEM with 2% FCS, 2 mM sodium butyrate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10 μg/ml gentamicin). Purity of eosinophils in the suspension was >98%, as determined by DiffQuick staining and viability was >95% by Trypan blue staining.
IMR32 nerve cell culture.
The IMR32 cell line was maintained at 37°C in an atmosphere of 5% CO2 in proliferation medium consisting of DMEM with 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10 μg/ml gentamicin. Upon reaching confluence, cells were plated at a density of 2 × 104/well on 24-well tissue culture plates and grown in proliferation medium. After 24 h, this medium was removed, and neurite outgrowth and a cholinergic phenotype were induced by the addition of differentiation medium. Cells were typically used after a further 6–8 days of differentiation in culture.
Measurement of EPO.
The release of EPO was measured as previously described (19), with some modifications. The incubation medium of IMR32 cells, differentiated as above, was replaced with differentiation medium containing 5 × 103 freshly isolated eosinophils/well. Eosinophils were also cultured alone (5 × 103 cells/well) to determine background levels of degranulation. After 2 and 24 h, aliquots of medium were removed from the cells and centrifuged at 300 g for 5 min. EPO was measured by an HVA oxidation assay (11) as follows. Supernatant aliquots (100 μl) were transferred to a 96-well plate and diluted with 100 μl 0.1 M glycine-NaOH buffer, pH 10.5, containing 0.3% (vol/vol) H2O2. The reaction was initiated by the addition of HVA (8 mM final concentration), and the samples were incubated for 1 h at 37°C. At the end of the incubation, fluorescence was measured in a microplate fluorescence reader (Bio-Tek FL600), with excitation filters set at 360 ± 40 nm and emission filters set at 485 ± 20 nm. Fluorescence values were converted to units of peroxidase activity by comparison with a peroxidase-conjugated immunoglobulin standard.
Measurement of LTC4.
Eosinophils were cultured either alone or with IMR32 cells, as for the measurement of EPO above. Aliquots of medium were removed at 2 and 24 h, centrifuged as above, and frozen at −80°C until required. Samples (50 μl) were transferred to a 96-well plate and assayed by EIA for LTC4, according to the manufacturer's instructions (Cayman Chemical). The concentration of LTC4 was calculated from a standard curve generated from standards provided in the EIA kit.
Detection of reactive oxygen species in IMR32 cells.
Nerve cells grown on 24-well tissue culture plates were loaded with 1 μM DHR in PBS for 30 min at 37°C and then washed two times in PBS before readdition of culture medium. DHR fluoresces when oxidized to rhodamine 123 (18). After incubation of IMR32 cells in the presence and absence of eosinophils, rhodamine 123 was detected in a microplate fluorescence reader (FL600; Bio-Tek), with excitation filters at 485 ± 20 nm and emission filters at 530 ± 20 nm. Rhodamine 123 fluorescence in the cells was also examined with a Zeiss Axiovert 35M fluorescence microscope fitted with a suitable optical filter set (451766).
Differentiated IMR32 cells were fixed with 4% (wt/vol) paraformaldehyde in PBS for 20 min, permeabilized with 0.5% (vol/vol) Triton X-100 in PBS for 10 min, and then incubated with blocking buffer [PBS, 5% (vol/vol) goat serum, and 0.3% (vol/vol) Triton X-100] for 1 h. The cells were then treated for 2 h at room temperature with either goat polyclonal antibody against the human p47phox (N-19) subunit of NADPH oxidase or goat IgG, both diluted in blocking buffer to a final concentration of 4 μg/ml. After three washes with PBS, the cells were then labeled with rabbit anti-goat IgG peroxidase conjugate (1 μg/ml) for 2 h. After three further PBS washes, immunostaining was detected with a peroxidase substrate kit (Vector) according to the manufacturer's instructions.
Eosinophil membrane preparations.
Immediately after isolation, eosinophils were resuspended in H2O, incubated for 15 min on ice, and then centrifuged at 1,500 g for 6 min. This process was repeated two more times, and then the resulting lysed cell membranes were resuspended in culture medium before addition to IMR32 cells of an amount equivalent to 5 × 103 whole eosinophils.
Measurement of ACh release from IMR32 cells.
Differentiated IMR32 nerve cells were radiolabeled with [methyl-3H]choline chloride (74 KBq/ml; sp act 20.8 GBq/mg choline) in Trowell's T8 medium containing 5% (vol/vol) FCS, 100 U/ml (wt/vol) penicillin, 100 μg/ml (wt/vol) streptomycin, and 10 μg/ml (wt/vol) gentamicin. The cells were washed with HEPES-buffered saline (HBS; 20 mM HEPES, 150 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl2, 0.8 mM MgSO4, and 25 mM glucose; pH 7.4) for 15 min. Hemicholinium-3 (10 μM) in HBS was then added for 15 min to inhibit reuptake of released choline. The cells were washed with HBS for 5 min before incubation of nerve cells with or without various amounts of human eosinophils in Trowell's T8 medium for 30 min. Each well was then field stimulated electrically (5 V; 1 Hz; 2 m/s pulse width) for 1 min, and 100-μl samples of medium were taken before and after stimulation. Each sample was centrifuged at 300 g for 1 min, and an aliquot was measured for released radioactivity in a liquid scintillation analyzer (Tri-Carb 1500; Packard, Pangbourne, UK).
Data represent means ± SE of at least three independent experiments. Statistical significance was estimated using one-way ANOVA with Dunnett's post hoc correction.
Eosinophil degranulation in the presence of IMR32 nerve cells.
Eosinophils cultured alone on tissue culture plates released detectable levels of EPO at 2 h, but there was an almost 10-fold increase in EPO release when eosinophils were cultured in the presence of IMR32 nerve cells for a similar time (Fig.1 A; P < 0.001). Similarly, after 24 h of coculture, the presence of nerve cells increased EPO release from eosinophils fivefold (Fig.1 B; P < 0.001). IMR32 cells themselves showed no EPO activity (data not shown). To determine whether the increased EPO release was dependent on adhesion, eosinophils were pretreated with either the very late antigen-4 (VLA-4) peptide inhibitor ZD-7349 (10 μM; see Ref. 12) or an MAb against CD11/18 (0.1 μg/ml) at concentrations previously shown to inhibit adhesion to IMR32 cells maximally by >80% (27) before addition to IMR32 cultures. After 2 h, ZD-7349 significantly reduced nerve-induced EPO release by 52 ± 12% (P < 0.01), the MAb to CD11/18 reduced EPO release by 55 ± 12% (P < 0.01), and a combination of the two treatments reduced release by 46 ± 15% (P < 0.01). Similar results were obtained after 24 h of coculture. IMR32 cells were also incubated with MAbs against ICAM-1 (2 μg/ml) and VCAM-1 (0.1 μg/ml; both concentrations have previously been shown to reduce eosinophil adhesion to IMR32 cells maximally by >80%). In these experiments, EPO release was reduced by 52 ± 6 and 47 ± 19% after 2 h of coculture (Fig. 1 A, bothP < 0.01). There was a similar inhibition of EPO release at 24 h (Fig. 1 B). Isotype-matched controls produced no significant reduction in EPO release.
Low baseline levels of LTC4 secretion were observed throughout a 24-h period from either IMR32 cells or eosinophils cultured alone. Coculture of IMR32 cells with eosinophils resulted in a significant enhancement of LTC4 secretion at both 2 and 24 h (Fig. 1, C and D, bothP < 0.001). Treatment of eosinophils with a combination of ZD-7349 and MAb against CD11/18 reduced LTC4secretion by 39 ± 18% (P < 0.05) at 2 h and 61 ± 12% (P < 0.01) at 24 h. Degranulation was also examined when eosinophils were cultured in inserts suspended above IMR32 cells (data not shown). Under these conditions, EPO release remained at or below baseline levels, further indicating that contact between eosinophils and nerves was necessary for the increase in EPO release.
fMLP-induced eosinophil degranulation in the presence of IMR32 nerve cells.
Both EPO secretion and LTC4 release were increased when eosinophils were treated with 10 nM fMLP for 2 or 24 h (Fig.2, A–D). The fMLP-induced release of EPO was significantly (P < 0.05) augmented in the presence of IMR32 cells after 2 h of incubation. The fMLP-induced increase in EPO appeared to be an additive effect of fMLP on eosinophils alone and on eosinophils and IMR32 cells together. Pretreatment of eosinophils with a combination of ZD-7349 and MAb against CD11/18 blocked only the nerve-augmented component of fMLP-induced EPO release (Fig. 2, A and B). Similar results were observed with LTC4 release (Fig. 2,C and D).
Eosinophil-induced generation of reactive oxygen species in IMR32 cells.
To determine whether active cellular pathways within IMR32 nerve cells or simply adhesion to the nerves was required for producing an increase in eosinophil degranulation, nerve cells were fixed with paraformaldehyde. Eosinophils adhered to paraformaldehyde-fixed IMR32 cells but failed to release EPO (data not shown). This led us to investigate the intracellular events stimulated in IMR32 cells after eosinophil adhesion. Interaction with VCAM-1 or ICAM-1 is known to stimulate the production of reactive oxygen species (ROS) in endothelial cells via NADPH oxidase (18). The production of ROS in IMR32 cells after adhesion of eosinophils was examined by loading the IMR32 cells with the reactive oxygen-sensitive fluorescent indicator DHR for 30 min before addition of eosinophils. Control nerves cultured alone showed low background rhodamine 123 staining when viewed with fluorescence microscopy (Fig.3 Ai), but when exposed to eosinophils for 2 h an intense green intracellular fluorescence was observed corresponding to the nerve cells (Fig. 3 A,iii and iv). This phenomenon was quantified in a microplate reader. There was a steady increase in ROS produced by nerves in contact with eosinophils over a 6-h time period, reaching a plateau at 24 h (Fig. 3 B). When compared with a standard curve constructed with hydrogen peroxide, the fluorescence indicated an approximate rate of production of 3 nmol ROS · h−1 · 100,000 cells−1. These increases were not because of ROS in eosinophils, since there was no significant increase in fluorescence when the eosinophils were loaded with DHR (Fig. 3 B).
When eosinophils were pretreated with MAb against CD18 (2 pg/ml), the production of ROS was reduced by 88 ± 4% at 2 h and 79 ± 10% at 24 h (P < 0.001). Similarly, anti-ICAM-1 MAb inhibited ROS generation at 2 and 24 h (Fig.4, A and B). Isotype-matched controls produced no significant reduction in ROS production (data not shown). Incubation with ZD-7349 had no significant effect on the production of ROS at 2 h but significantly reduced the levels at 24 h (Fig. 4, A and B). Eosinophils separated from the nerves by inserts failed to induce neuronal ROS production. IMR32 cells were pretreated for 30 min with NADPH oxidase inhibitors, either apocynin (5 mM) or DPI (1 μM). Apocynin reduced the eosinophil-mediated ROS production by 83 ± 6% at 2 h and 33 ± 16% after 24 h of coculture, whereas DPI completely blocked ROS production at both time points (Fig. 4, C and D). In contrast, the nitric oxide synthase (NOS) inhibitorN G-monomethyl-l-arginine (l-NMMA; 1 mM) significantly reduced the levels of ROS after 24 h but not after 2 h of coculture.
To determine that the NADPH oxidase was localized to the IMR32 nerve cells, we examined the expression of the p47phox subunit by immunocytochemistry. Minimal background staining was observed when the nerves were treated with the primary antibody, goat anti-human p47phox (N-19) polyclonal antibody alone (Fig.5 A), the secondary antibody, rabbit anti-goat horseradish peroxidase conjugate alone (Fig.5 B), or the isotype-matched control goat IgG in combination with secondary antibody (Fig. 5 C). An intense blue-gray staining was observed when the cells were exposed to both primary and secondary antibody (Fig. 5 D), indicating the presence of the p47phox subunit of NADPH oxidase in IMR32 cells.
Eosinophil degranulation induced by neuronal ROS.
Eosinophils cultured with IMR32 cells treated with DPI released 65 ± 26% (P < 0.05) and 49 ± 2% (P < 0.01) less EPO than eosinophils cultured in the presence of control nerves at 2 and 24 h, respectively (Fig.6, A and B). The specific NADPH oxidase inhibitor apocynin reduced degranulation to a similar extent, but the specific NOS inhibitor l-NMMA had no effect on degranulation (Fig. 6, A and B), suggesting that the NADPH oxidase enzyme and its reaction product superoxide are important mediators of eosinophil degranulation. Complete inhibition of ROS in nerves did not prevent all nerve-induced eosinophil degranulation, suggesting that there were other neural factors responsible.
To determine if the ROS (or other neural factors) caused degranulation of only those eosinophils directly in contact with nerves, we separated eosinophils from nerves in a Transwell insert and then generated ROS by incubation of nerves with lysed eosinophil membranes. Under these conditions, there was an increase in neuronal ROS comparable to the addition of whole eosinophils that was also sensitive to anti-CD18 MAb (data not shown). Furthermore, there was a significant release of EPO from nonadhered eosinophils at 2 and 24 h, and this degranulation was completely blocked by DPI (Fig. 6, C and D), indicating that ROS from nerves could mediate degranulation of eosinophils not directly in contact with nerves.
Effect of eosinophils on stimulated ACh release.
Eosinophils can elicit M2 receptor dysfunction in vivo (5) and thereby generate increased levels of ACh release. To determine whether the in vitro interactions we observed between nerves and eosinophils might influence this function of nerve cells, the response of IMR32 cells to electrical stimulation in the presence of eosinophils was examined. IMR32 cells release ACh when electrically stimulated, and this release is reduced by M2 receptor agonist. When IMR32 cells were stimulated in the presence of an increasing number of eosinophils, there was a significant increase in the amount of ACh released (Fig. 7,P < 0.05), suggesting that nerve-induced eosinophil degranulation might contribute to neuronal dysfunction. The increase in ACh release was inhibited by preventing eosinophil adhesion with an antibody to CD11/18.
The results of this study show that contact between eosinophils and cholinergic nerves leads to eosinophil degranulation and an increase in ACh release from the nerves. The nerve-induced eosinophil degranulation was dependent on the interactions of the eosinophil integrins CD11/18 and VLA-4 and the neuronal adhesion molecules ICAM-1 and VCAM-1. In particular, adhesion to IMR32 cells resulted in the production of neural ROS, which, along with other factors released from the nerves, induced the eosinophil degranulation. Similarly, the increased ACh release was dependent on eosinophil adhesion to the nerves.
In previous in vivo and in vitro studies, we have shown that eosinophils influence cholinergic nerve function (5, 27). To investigate the nature of the interaction between these two cell types, we first investigated the effect of contact with cholinergic IMR32 cells on eosinophil function. Both eosinophil degranulation, as indicated by the release of EPO, and eosinophil activation, as indicated by the production of LTC4, were increased on incubating the cells together. Both of these effects occurred within 2 h and continued beyond that time, indicating an active continuous process of degranulation. The eosinophil degranulation required adhesion to live nerves, since it was completely inhibited by the use of Transwell inserts to prevent eosinophil contact with nerves or alternatively by paraformaldehyde fixation of the nerves.
Previous reports have indicated that eosinophil adhesion and degranulation are closely linked. For example, stimulation of eosinophil degranulation with chemotactic factors and cytokines such as platelet-activating factor (PAF) and granulocyte macrophage colony-stimulating factor is mediated by β2-integrins (14). Studies of eosinophils in coculture with an epithelial cell line have indicated that adhesion is necessary for the release of eosinophil cationic protein (ECP; see Ref. 28). In that study, the eosinophils were found to release ∼15% of total ECP, and in our nerve-induced degranulation studies we found a similar level of total EPO release. Furthermore, the levels of released peroxidase activity that we observed were of a similar magnitude to that induced by PAF (16), suggesting the effect to be physiologically significant. We have previously shown that eosinophil adhesion to IMR32 nerve cells involves the adhesion molecules VLA-4 and CD11/18 interacting with IMR32 cell VCAM-1 and ICAM-1. Inhibition of either adhesion molecule prevents eosinophil-nerve adhesion (27); this is consistent with a process of inside-to-outside signaling, strengthening adhesion when eosinophils encounter both adhesion molecules. Prevention of adhesion with antibodies to ICAM-1 or VCAM-1 or their ligands either alone or in combination significantly, but not fully, inhibited the degranulation.
Other work has shown that adhesion of eosinophils to other cells is insufficient to induce degranulation but rather that adhesion leads to priming, as demonstrated by enhanced responses to the secretogogue fMLP (21). We investigated whether adhesion to IMR32 nerve cells also led to eosinophil priming. Adhesion for 2 or 24 h enhanced the release of EPO and generation of LTC4 in response to fMLP. This priming effect of the nerve cells was completely blocked with either an MAb against CD11/18 or by the peptide inhibitor ZD-7349 against VLA-4, suggesting that the priming was the result of adhesion molecule-induced signaling in the eosinophils. Thus stimulation by IMR32 cells of eosinophil degranulation appeared to be mediated by a different mechanism from that responsible for priming of secretory responses to exogenous factors such as fMLP.
Having demonstrated that adhesion leads to both degranulation and priming of eosinophils and that these are mediated by different processes, we investigated the underlying mechanisms in more detail, addressing specifically the role of adhesion molecules in generating signals in nerves. Using a fluorescent probe system that is highly sensitive to changes in cellular reactive oxidants, we assessed the effects of adhesion of eosinophils to nerves on the generation of intraneuronal ROS. After contact with eosinophils, ROS were generated in nerves within half an hour and persisted for a period in excess of 24 h. This was inhibited by antibodies to either CD11/18 or ICAM-1, suggesting that these interactions were responsible for the generation of neuronal ROS and is similar to observations in neutrophil-endothelial cell interactions (31, 32). Furthermore, the production of neuronal ROS was blocked by both DPI and apocynin, which are inhibitors of NADPH oxidase. After 24 h, the NOS inhibitor l-NMMA also reduced the production of ROS, suggesting that a secondary generation of peroxynitrite also occurred (20).
The expression of NADPH oxidase was previously believed to be restricted to nonneuronal cells. However, recent studies have indicated that this enzyme is present in sympathetic neurones (29). We have now also shown by immunocytochemistry that IMR32 cells express the p47phox subunit of NADPH oxidase. The NADPH oxidase complex has been studied extensively in inflammatory cells, including eosinophils and neutrophils, where it functions to produce high levels of ROS. In our experiments, however, we could not detect the production of ROS within eosinophils, but rather the ROS were derived from the nerves alone. The levels of ROS were comparable to those previously described by activation of NADPH oxidase in inflammatory cells (17). It was clearly important to identify if the ROS generation within the cholinergic IMR32 cells was responsible for the nerve-induced eosinophil degranulation. Eosinophils were therefore cocultured with IMR32 cells in the presence of inhibitors of NADPH oxidase, and the effect of these on degranulation was assessed. When eosinophils were added to IMR32 cells that had been pretreated with DPI or apocynin but not l-NMMA and therefore generated no ROS, there was a significant but incomplete inhibition of eosinophil degranulation. These data suggest that eosinophil adhesion to IMR32 cells leads to degranulation only partly through the generation of ROS. There may be a direct effect of ROS released from the nerves on eosinophil function. For example, the ROS may alter cell adhesion, or the production of eosinophil produced factors such as platelet-activating factor (22, 25, 26).
The fact that inhibition of the neuronal ROS did not completely inhibit the nerve-induced degranulation suggests that eosinophil adhesion might lead to another neural factor being released. Experiments were performed using eosinophil membranes as a means of cross-linking ICAM-1 with CD11/18 but independently of local EPO release, since it was necessary to determine whether degranulation could occur in eosinophils not in contact with nerves. When eosinophils were separated from nerves stimulated with eosinophil membranes, there was some degranulation but less than what occurred when whole eosinophils were in contact with nerves. This reduction in degranulation may have a number of explanations. Either a nerve-released factor did not pass through the Transwell membrane or eosinophils needed to be in contact with nerves for them to respond to this factor. Alternatively, eosinophil-nerve contact leads to release of an autocrine mediator from eosinophils that contributes to degranulation. For instance, other studies have shown that PAF generated by eosinophils that have adhered via β2-integrins can directly induce eosinophil degranulation (1). Our preliminary experiments (results not shown) indicate that the factor in our preparations was not PAF, since the PAF antagonist CV-6209 did not affect eosinophil degranulation. Further studies will be required to identify the nature of the other neural factors that induce eosinophil activation and degranulation.
Having demonstrated that eosinophil degranulation occurred after contact with cholinergic nerves, we assessed the effects of this process on stimulated ACh release. In experiments in which eosinophils were allowed to adhere to IMR32 cells, both spontaneous and stimulated ACh release was significantly enhanced compared with baseline levels. This increase in ACh release was significantly inhibited by pretreatment with an antibody to CD11/18. Thus eosinophil adhesion to IMR32 cells was associated with enhanced cholinergic nerve activity. This is consistent with the eosinophils influencing activity at M2 receptors.
In summary, these studies indicate that contact between eosinophils and cholinergic nerves leads to eosinophil degranulation resulting from neural factors, including ROS. Eosinophil contact with nerves also leads to an increase in ACh release from the nerves. Eosinophil degranulation in association with nerves and the subsequent alteration in nerve function may occur in a number of disease states. For example, degranulation of eosinophils close to airway nerves is seen in animal models of antigen-induced hyperreactivity and in gastrointestinal diseases (13). In addition, eosinophil proteins are implicated in a number of neurological conditions, including Gordon's syndrome, muscle rigidity, and ataxia resulting from eosinophil-derived neurotoxin (7) and also as part of the immune response of some brain cancers (23, 30, 33). Hence the mechanisms of nerve-induced eosinophil degranulation described in this report may be relevant to a number of neuroinflammatory conditions.
This work was supported by the Wellcome Trust (059037), the British Lung Foundation (PD97/5), and by a Samuel Crossley- Barnes studentship (to D. A. Sawatzky).
Address for reprint requests and other correspondence: R. W. Costello, Dept. of Medicine, RCSI, Beaumont Hospital, Dublin 9, Ireland (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published January 25, 2002;10.1152/ajplung.00278.2001
- Copyright © 2002 the American Physiological Society