Abstract
Eosinophils release a number of mediators that are potentially toxic to nerve cells. However, in a number of inflammatory conditions, such as asthma and inflammatory bowel disease, it has been shown that eosinophils localize to nerves, and this is associated with enhanced nerve activity. In in vitro studies, we have shown that eosinophil adhesion via neuronal ICAM-1 leads to activation of neuronal NF-κB via an ERK1/2-dependent pathway. In this study, we tested the hypothesis that eosinophil adhesion to nerves promotes neural survival by protection from inflammation-associated apoptosis. Exposure of differentiated IMR-32 cholinergic nerve cells to IL-1β, TNF-α, and IFN-γ, or culture in serum-deprived medium, induced neuronal apoptosis, as detected by annexin V staining, caspase-3 activation, and DNA laddering. Addition of human eosinophils to IMR-32 nerve cells completely prevented all these features of apoptosis. The mechanism of protection by eosinophils was by an adhesion-dependent activation of ERK1/2, which led to the induced expression of the antiapoptotic gene bfl-1. Adhesion to nerve cells did not influence the expression of the related genes bax and bad. Thus, prevention of apoptosis by eosinophils may be a mechanism by which these cells regulate neural plasticity in the peripheral nervous system.
The airway and enteric smooth muscle and mucus glands are innervated with a wide variety of motor and sensory nerves that maintain normal function of these structures (1). Increased activity of some of these nerves is seen in conditions such as asthma and rhinitis after exposure to allergens, pollutants, and viruses or during exacerbations of inflammatory bowel disease or enteric infections (2, 3, 4). Increased neural activity leads to enhanced smooth muscle contraction and mucus production, both of which are part of the host’s defense responses.
In all of these inflammatory conditions, we and others have shown that eosinophils localize to, and release granular proteins in association with, airway and enteric nerves (5, 6). Eosinophils, in particular, release a number of compounds that are toxic to nerves, such as reactive oxygen species, proteases, and the eosinophil cationic proteins, major basic protein and eosinophil-derived neurotoxin (7, 8). Thus, it may be expected that eosinophil localization may be harmful for nerves. However, in animal models, eosinophils directly influence a variety of aspects of neural function, including nerve remodeling, survival, and neurotransmitter release (5, 6).
In in vitro studies, we have shown that eosinophils adhere to cholinergic nerves via specific adhesion molecules (9). Eosinophil adhesion leads to the activation of a number of intracellular pathways within the nerve cells, including the MAPKs ERK1/2 and p38, and in turn these lead to activation of NF-κB and AP-1, respectively (10). The nuclear transcription factor NF-κB has been shown, in certain circumstances, to have a neuroprotective effect (11). Thus, we hypothesized that in addition to activating the nerves, eosinophil adhesion-induced activation of NF-κB may have a beneficial effect for the nerves, providing protection from the hazardous inflammatory products.
To address this hypothesis, we used a coculture system of a cholinergic nerve cell line, IMR-32 cells; these were cultured with eosinophils in the presence of stimuli that induce neuronal apoptosis. To distinguish adhesion-dependent effects from factors released from the eosinophils, we used eosinophil membranes for these studies. Eosinophil membranes adhere to nerve cells, but do not contain nuclear or cytoplasmic contents.
Materials and Methods
Materials
DMEM Plus glutamax, FCS, penicillin/streptomycin solution was purchased from Invitrogen Life Technologies (Paisley, U.K.). The IMR-32 cell line was obtained from European Cell Culture Collection (Salisbury, U.K.) and depleted of fibroblasts with immunomagenetic antifibroblast microbeads and LD MACS separation columns (Miltenyi Biotec, Bisley, U.K.). Gentamicin, trypan blue, Tri-reagent, and all common buffer constituents were obtained from Sigma-Aldrich (Poole, U.K.). All cell culture plastic materials were obtained from Cuminn (Dublin, Ireland). Speedy-Diff was obtained from Clin-Tech (Clacton on sea, U.K.); Ficoll-Paque Plus was purchased from Amersham Biosciences (Little Chalfont, U.K.). CD16 microbeads and MACS VS+ columns were purchased from Miltenyi Biotec. The ERK1/2 inhibitor PD98059 was purchased from Cell Signaling Technology (Beverly, MA). Mouse anti-human ICAM-1 mAb (IgG1, clone15.2) and mouse anti-human VCAM-1 mAb (IgG1, clone 1.G11B1) were purchased from Autogen Bioclear (Calne, U.K.), and SB239063 was a gift from M. Salmon (GlaxoSmithKline, Philadelphia, PA). Purified eosinophil proteins were a gift from G. Gleich (University of Utah, Salt Lake City, UT), prepared as previously described. Mouse anti-human major basic protein (MBP)3 mAb was purchased from Monosan (Uden, The Netherlands); goat polyclonal anti-human MBP and eosinophil peroxidase (EPO) Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA); mouse monoclonal anti-human eosinophil-derived neurotoxin (EDN) was obtained from Research Diagnostics (Flanders, NJ). Secondary HRP-conjugated Abs and Immunoglo reagents A and B were purchased from New England Biolabs (Beverley, MA). The Annexin VFITC⇓. Reagents for cDNA synthesis and LightCycler-FastStart DNA master SYBR Green 1 were from Roche Diagnostics (Mannheim, Germany). Suicide-Track DNA ladder isolation kit and the fluorogenic caspase-3 substrate, Ac-DEVD-AMC, were purchased from Merck Bioscience (Nottingham, U.K.).
Primers used for gene amplification
IMR-32 nerve cell culture
The human cholinergic neuroblastoma cell line IMR-32 was depleted of fibroblasts, as described previously (10). The cells were maintained in proliferation medium (DMEM Plus glutamax, 5% FCS, 100 U/ml penicillin/streptomycin, 10 μg/ml gentamicin) at 37°C in an atmosphere of 5% CO2 until the cells were confluent. The cells were then plated at a density of 5 × 105/well in six-well cell culture dishes and grown in proliferation medium for a further 48 h when the proliferation medium was then replaced by differentiation medium (DMEM Plus glutamax, 2% FCS, 2 mM sodium butyrate, 100 U/ml penicillin/streptomycin, 10 μg/ml gentamicin). The cells were used for experimentation after a further 6–7 days of differentiation in culture. Cells cultured in this manner have a cholinergic phenotype; they express functional M2 receptors and release acetylcholine in response to electrical stimulation, and so are a useful model to study cholinergic cells in vitro (12, 13, 14, 15).
Eosinophil isolation
Eosinophils were prepared from 45 ml of peripheral blood drawn from healthy human volunteers. After phlebotomy, 15 ml of blood was added to 25 ml of PBS to which 100 U of heparin had been added, and 30 ml of this blood/PBS mixture was then layered over 23 ml of Ficoll (1.077 ± 0.001 g/ml) and centrifuged at 720 × g for 20 min at room temperature. The upper layer of serum and mononuclear cells was discarded, and the pellet-containing granulocytes and erythrocytes were subjected to hypotonic lysis. The granulocytes were then resuspended in MACS buffer (PBS with 2 mM EDTA and 0.5% BSA) with anti-CD16 immunomagnetic beads and passed through a magnetic separation column. The eluted eosinophils were collected and resuspended in differentiation medium, and their viability and purity were determined by trypan blue and Speedy-Diff staining. Only populations of eosinophils >98% pure and >95% viable were used in coculture experiments.
Preparation of inactivated eosinophils
To distinguish contact-dependent effects from factors released from the eosinophils, eosinophils were resuspended immediately after isolation in sterile, deionized water, incubated on ice for 15 min, and then centrifuged at 10,000 × g for 10 min at 4°C. This process was repeated twice, and the resulting lysed cell membranes were resuspended in differentiation medium before addition to IMR-32 cells at a concentration equivalent to 2 × 105 whole eosinophils/well. Alternatively, eosinophils were inactivated with 4% paraformaldehyde for 20 min, centrifuged at 400 × g for 5 min, and suspended in culture medium, and the process was repeated twice more. Flow cytometry and functional studies showed that these preparations retain an ability to adhere via CD11/18 and VLA-4 to nerve cells (9, 10, 14).
Coculture experiments
Differentiated cholinergic IMR-32 cells were grown to confluence in six-well plates for 6–7 days. Cells were then cultured in serum-free medium or treated with a cytomix containing TNF-α (10 ng/ml), IL-1β (10 ng/ml), and IFN-γ (25 ng/ml), concentrations previously shown to induce apoptosis (16). IMR-32 cells were then coincubated with whole eosinophils or eosinophil membranes for various times. In experiments designed to investigate the role of MAPKs and adhesion, cells were pretreated for 2 h with the ERK1/2 inhibitor PD98059 (50 μM), the p38 kinase inhibitor SB239086 (10 μM), or anti-ICAM-1 (0.25 μg/ml) and VCAM-1 (0.25 μg/ml) Abs.
Annexin V-binding apoptosis assay
Differentiated cholinergic IMR-32 were cocultured with whole eosinophils or inactivated eosinophils for various times. Medium and nonadherent cells were removed, and cells were harvested in PBS and washed twice by centrifugation at 500 × g. Cells were then incubated with Annexin VFITC and propidium iodide (PI), according to manufacturers’ guidelines. The cells were subsequently analyzed by flow cytometry (Coulter Epics XL; Beckman Coulter, High Wycombe, U.K.). Data from 10,000 events were collected in logarithmic mode. Dual staining with Annexin VFITC and PI allowed for the differentiation between early apoptotic cells (Annexin VFITC positive), late apoptotic and/or necrotic cells (Annexin VFITC and PI-positive), and viable cells (unstained).
RNA extraction and RT-PCR
After removal of medium, IMR-32 nerve cells were harvested and washed in warm PBS, lysed at room temperature in Tri-reagent, and RNA extracted, according to the manufacturer’s guidelines. For both quantitative LightCycler PCRs and semiquantitative RT-PCRs, 1 μg of total RNA was reverse transcribed into cDNA using an oligo(dT)15 primer from the first strand cDNA synthesis system (Roche Diagnostics), according to the manufacturer’s instructions. For quantitative PCR, amplification of cDNA was conducted on the LightCycler (Roche Diagnostics) using Fast Start TaqDNA polymerase containing the dsDNA-binding dye SYBR Green 1. The samples were continuously monitored during the PCR, and fluorescence was acquired every 0.1°C. PCR mixtures contained 0.5 μM of either β-actin, Bfl-1, Bad, or Bid primer pairs (Table I⇑). The samples were denatured at 95°C for 10 min, followed by 45 cycles of annealing and extension at 95°C for 12 s, 55°C for 5 s, and 72°C for 8 s. Characteristic melting curves were obtained at the end of amplification by cooling the samples to 65°C for 15 s, followed by further cooling to 40°C for 30 s. Serial 10-fold dilutions were prepared from known quantities of β-actin and M2 PCR products, which were then used as standards to plot against the unknown samples. Quantification of data was analyzed using the LightCycler analysis software, and values were normalized to the level of β-actin expression for each sample on the same template cDNA.
For semiquantitative PCR, RT-PCR analysis of cDNA preparations was conducted in 50 μl reactions with Taq-DNA polymerase, and the primer sets outlined in Table I⇑. PCR products were separated by 2% agarose gel electrophoresis and photographed under UV illumination. Band intensities were quantified by laser densitometry scanning. The results were expressed as a ratio of the band intensity relative to the corresponding β-actin band obtained by amplification of the same template cDNA.
Western blotting
Protein extracts from eosinophil membranes (10 μg) or purified eosinophil proteins (100 ng–1 μg) were heated to 95°C in sample buffer (100 mM Tris, pH 6.8, 2% (w/v) SDS, 0.002% (w/v) bromphenol blue, 20% (v/v) glycerol, 10% (v/v) 2-ME) and separated by SDS-PAGE on 10% polyacrylamide-separating gel overlaid with 4% stacking gel at 500 V for 1 h. The separated proteins were transferred on to nitrocellulose membranes in transfer buffer (20 mM Tris, 150 mM glycine, 0.01% (w/v) SDS, 20% (v/v) methanol) at 500 V overnight. For immunodetection, membranes were incubated in blocking buffer (Dulbecco’s PBS containing 0.2% (w/v) I-block and 0.1% (v/v) Tween 20) for 1 h at room temperature, then incubated for 2 h in blocking buffer containing the individual respective Ab (1:500 for each). After six 5-min washes in washing buffer (PBS, pH 7.4, 0.1% (v/v) Tween 20), membranes were incubated for 1 h in blocking buffer containing anti-mouse IgG HRP conjugate (1/2000).
Following further washes, blots were incubated with an HRP substrate prepared from Lumiglo Reagent A and Peroxide Reagent B (both Cell Signaling Technology) for 30 s. Blots were then exposed to X-OMAT light-sensitive film to obtain an image.
Caspase-3 assay
IMR-32 cells were seeded onto six-well plates at a density of 1 × 105 cells/well and induced to differentiate for 4–6 days. The nerve cells were then treated with cytomix (as above) or switched to serum-free medium for a further period of 48 h in the absence or presence of inactivated eosinophils prepared from 1 × 105 live eosinophils. The nerves were then gently scraped into cell culture medium and centrifuged at 200 × g for 5 min. The pellet containing both the adherent and detached apoptotic cell fractions was then treated with lysis buffer (1% Triton X-100, 130 mM NaCl, 10 mM Tris-HCl, 10 mM NaH2PO4, pH 7.5) and incubated on ice for 15 min. Nonsolubilized protein was removed by centrifugation at 12,000 × g for 15 min at 4°C, and the supernatant then combined with an equal volume of caspase reaction buffer containing 40 mM HEPES (pH 7.5), 20% glycerol, 4 mM DTT, and 40 μM caspase-3 substrate (Ac-DEVD-AMC). Fluorescence generated by cleavage of the AMC moiety was measured after 2 h in a Bio-Tek FL600 plate reader (Bio-Tek Instruments, Watford Herts, U.K.) with excitation filters at 360 ± 40 nm and emission filters at 485 ± 20 nm. Caspase-3 activity was expressed as unit fluorescence per μg protein after determination of protein levels with the Bio-Rad (Hercules, CA) DC protein assay kit.
DNA laddering
IMR-32 cells were cultured as for caspase assays, but treated for a period of 72 h with either cytomix or serum-free medium. Apoptotic DNA was then isolated according to the manufacturer’s guidelines with the Suicide-Track DNA laddering kit (Oncogene Research Products, San Diego, CA). Briefly, this involved gently scraping the cells into medium and subsequent centrifugation at 200 × g for 5 min. The pellet containing both the adherent and detached apoptotic cell fractions was then treated with extraction buffer designed to isolate apoptotic DNA fragments from high m.w. chromatin. DNA was then precipitated with 3 M NaOH and 2-propanol. The samples were then loaded onto a 1.5% (w/v) agarose gel and electrophoresed for 4 h at constant 50 V. DNA was visualized by UV illumination, recorded, and then analyzed densitometrically on an Alpha Innotech (San Leandro, CA) imaging system.
Statistical analysis
Results are expressed as mean ± SEM, unless otherwise indicated. Data were compared using ANOVA; a p value of <0.05 was considered significant.
Results
Eosinophil coculture protects IMR-32 cholinergic nerve cells from apoptosis
IMR-32 cells were grown to confluence, and both neurite outgrowth and a cholinergic phenotype were induced by differentiation medium, as described previously (10). This medium was replaced with differentiation medium containing TNF-α (10 ng/ml), IL-1β (10 ng/ml), and IFN-γ (25 ng/ml) for a further 16 h. The IMR-32 cells were then stained for annexin V and PI and subjected to FACS analysis. Compared with control cells, cytokine treatment induced a 4- to 5-fold increase in neuronal apoptosis (see Fig. 1⇓, A, B, and D). Coculture of eosinophils with IMR-32 cells, at the same time as the cytokine exposure, led to inhibition of the cytokine-induced apoptosis (see Fig. 1⇓, C and D). Inactivated eosinophil membrane preparations conferred a similar degree of protection from apoptosis as freshly isolated intact eosinophils (Fig. 1⇓D). Serum-deprived IMR-32 cells demonstrated a 5-fold increase in annexin staining, but coculture with eosinophil membrane preparations (1 × 104-1 × 106 eosinophil/ml) at the same time prevented the apoptotic effect of serum deprivation in a concentration-dependent manner (see Fig. 1⇓, E and F). In contrast, neutrophils and neutrophil membrane preparations at a range of concentrations (1 × 104-1 × 106 neutrophil/ml) did not protect IMR-32 cells from apoptosis induced by either an inflammatory cytomix or serum deprivation (Fig. 2⇓A).
Eosinophils or eosinophil membranes protect IMR-32 cholinergic nerve cells from apoptosis induced by both cytokines or serum deprivation. Coculture experiments were performed, as described; IMR-32 nerve cells were harvested at 16 h for survival analysis. A cytomix (B–D) of TNF-α (10 ng/ml), IL-1β (10 ng/ml), and IFN-γ (25 ng/ml) or serum deprivation (E and F) was used to induce apoptosis. FACS scattergrams from representative experiments are shown in A–C; dual staining with PI/Annexin VFITC was used to distinguish apoptotic, necrotic, and live cells. D, IMR-32 cells were cocultured in presence of eosinophils or eosinophil membranes alone; graph shows mean ± SEM for four independent experiments. E, Eosinophil membranes protect IMR-32 nerve cells from apoptosis induced by serum deprivation. F, IMR-32 cells were cocultured in presence of indicated number of eosinophils to generate a dose-response curve for protection from serum deprivation-induced apoptosis. ∗, p < 0.05; ∗∗, p < 0.005.
A, Neutrophils do not protect IMR-32 nerve cells from apoptosis. Differentiated cholinergic nerve cells were coculture with neutrophils at 1 × 105 for 24 h in the presence or absence of cytokines (TNF-α, 10 ng/ml; IL-1β, 10 ng/ml; and IFN-γ, 25 ng/ml) or serum deprivation (SD) to induce apoptosis. Cells were stained with PI and annexin V with FACS analysis. Percentage of apoptotic cells represents those cells staining with annexin V only. n = 3; ∗, p < 0.05. B, Eosinophil membranes do not contain EPO nor EDN. Western blots against the indicated amounts of standard purified protein, total eosinophil protein, or eosinophil membrane protein are shown. C, Eosinophil membranes do not contain RNA; 1% agarose gel electrophoresis to demonstrate presence of RNA is shown.
Investigations to characterize the eosinophil membranes showed that they did not contain RNA (Fig. 2⇑C) and did not express the eosinophil cationic proteins EPO, EDN by Western blot analysis (Fig. 2⇑B), nor MBP within the limits of detection of the available Abs.
Effect of eosinophils on caspase-3 activity and DNA fragmentation
Apoptosis commonly involves the activation of caspases downstream of the early membrane changes that we have detected by annexin V staining (17). Thus, we examined the level of caspase-3 activity in nerve cell lysates after exposure to cytokines and serum-free medium. In these studies, cytokine treatment resulted in a small, but statistically significant increase in caspase-3 activation in the nerve cells at 48 h (172.3 ± 33.83 fluorescence/μg protein at baseline vs 243.1 ± 31.3, at 48 h; p = 0.0034). There was a much larger increase (105.3 ± 23.64 fluorescence/μg protein vs 516.7 ± 101.6, p = 0.013) in activation of caspase-3 in response to serum deprivation for 72 h. There was a significant attenuation of the elevated caspase-3 activity when the nerves were incubated in the presence of inactivated eosinophil membrane preparations (Fig. 3⇓, p = 0.005).
Eosinophils significantly inhibit serum deprivation-induced caspase-3 activity in IMR-32 nerve cells. Differentiated IMR-32 cells were switched to serum-deprived medium (SD) for a period of 48 h in the absence or presence of inactivated eosinophils (eos) prepared from 1 × 105 live eosinophils. Cell lysates were then prepared, as in Materials and Methods, and incubated with 40 μM caspase-3 substrate (Ac-DEVD-AMC) for 2 h. Fluorescence was recorded in a plate reader with excitation filters at 360 ± 40 nm and emission filters at 485 ± 20 nm. Caspase-3 activity was determined as unit of fluorescence per mg protein, and is expressed as fold increase above untreated cells. n = 5 independent experiments; ∗∗, p < 0.01 compared with serum deprivation alone.
To further confirm that eosinophils or eosinophil membrane preparations prevented neuronal apoptosis, we isolated DNA from either cytomix-treated or serum-deprived IMR-32 cells and examined this by agarose gel electrophoresis. Both treatments gave rise to a DNA ladder pattern characteristic of apoptosis (Fig. 4⇓A). Coincubation with eosinophil membrane preparations significantly (p < 0.05) blocked the formation of DNA fragments induced by cytomix treatment and reduced it by 67% after serum deprivation.
Inactivated eosinophils protect IMR-32 nerve cells from DNA laddering induced by either cytokines or serum deprivation. A, Differentiated IMR-32 cells were either left untreated (con) or incubated for a period of 72 h with either cytomix or serum-deprived (SD) medium in the absence or presence of inactivated eosinophils (eos) prepared from 1 × 105 live eosinophils. Apoptotic DNA was then isolated, as described in Materials and Methods, loaded onto a 1.5% (w/v) agarose gel, electrophoresed for 4 h at constant 50 V, and visualized by UV illumination. Numbers indicate size of DNA fragments in base pairs. B, Gels were analyzed using an Alpha Innotech densitometry imaging system to obtain an arbitrary integrated density value. Results are the mean ± SEM from four independent experiments; ∗, p < 0.05 significantly different from treatment in the absence of eosinophils.
Dependence of eosinophil protection from apoptosis on adhesion and ERK1/2 signaling
We subsequently studied which eosinophil-activated signaling pathways might protect against nerve apoptosis. The protective effect of inactivated eosinophil preparations on cytokine-induced neuronal apoptosis was completely inhibited by pretreatment with blocking Abs to prevent eosinophil CD11/18 and VLA-4 interactions with their counter ligands on nerves, ICAM-1 and VCAM-1, respectively (Fig. 5⇓). Next, we examined the role of the MAPKs ERK1/2 and p38 in conveying protection from apoptosis within the nerves because these are activated by ligation of ICAM-1 and VCAM-1, respectively (10). Pretreatment of the nerve cells with the ERK1/2 inhibitor PD98059 (50 μM) completely blocked the protective effect of eosinophil membranes on neuronal apoptosis (p = 0.01). In contrast, pretreatment with SB239063 (10 μM), an inhibitor of p38, had no effect on eosinophil protection from neuronal apoptosis (Fig. 5⇓). These findings indicate that eosinophils confer resistance against apoptosis of IMR-32 nerve cells in a manner that is dependent on adhesion-induced activation of intraneuronal ERK1/2.
Adhesion of eosinophils to IMR-32 cells and the activation of ERK1/2 are required for protection from cytokine-induced apoptosis. IMR-32 nerve cells were cultured in the presence or absence of inactivated eosinophil membranes, cells were harvested at 16 h, and apoptosis was detected by double staining with PI/annexin V. Coculture was conducted with or without pretreatment for 2 h with the ERK1/2 inhibitor PD98059 (50 μM), the p38 kinase inhibitor SB23063 (10 μM), or anti-ICAM-1 and anti-VCAM-1 (0.25 μg/ml) Abs. Mean ± SEM of four independent experiments are shown. ∗∗, p < 0.005.
Eosinophils stimulate expression of Bfl-1 in IMR-32 cholinergic nerve cells
In additional experiments designed to investigate the potential mechanism of eosinophil protection from cytokine-induced apoptosis, we assessed changes in expression of the apoptosis-associated genes Bfl-1, Bid, and Bad. Control IMR-32 cells, grown in normal differentiation medium, expressed low levels of Bfl-1, and this was reduced after 16 h of cytomix treatment to 28 ± 0.21% of baseline (p < 0.01; Fig. 6⇓, A and B). Coculture of IMR-32 cells with eosinophil membranes increased Bfl-1 expression in cytokine-treated nerve cells at 16 h to 165 ± 0.92% of baseline (p < 0.01, n = 4). This increase in Bfl-1 expression was first evident at 2 h, and it was sustained for at least 24 h. Neither cytomix nor eosinophils altered the expression of the proapoptotic genes Bid and Bad in IMR-32 cells (Fig. 6⇓, A and D). Pretreatment with PD98059 prevented the eosinophil-induced Bfl-1 expression. Similarly, eosinophil-induced Bfl-1 expression was attenuated when adhesion was prevented with inhibitors of ICAM-1 and VCAM-1 (Fig. 6⇓B). Equivalent results were found when eosinophil membrane preparations were cocultured with serum-deprived nerve cells (Fig. 6⇓C). Thus, when serum-free medium was used as another means of inducing apoptosis, a similar protection by eosinophils against neuronal apoptosis by an adhesion-dependent pathway could be demonstrated.
Eosinophil membranes stimulate the transcription of the prosurvival gene Bfl-1 in cytokine-treated and serum-deprived IMR-32 cells. IMR-32 nerve cells were cocultured with eosinophil membranes, as described; some cells were pretreated for 2 h with the ERK1/2 inhibitor PD98059 or anti-ICAM-1 Ab. A, Up-regulation of Bfl-1 mRNA in IMR-32 cells by eosinophil membranes is shown. Transcription of the proapoptotic genes Bid and Bad is unaffected (A and D). B, Results of four independent experiments showing mean ± SEM; Bfl-1 gene up-regulation in IMR-32 cells is blocked by pretreatment with the ERK1/2 or anti-ICAM-1. ∗∗, p < 0.005. C, Equivalent results in serum-deprived cells with Bfl-1 up-regulation in nerve cells by eosinophil membranes.
Discussion
The results of these studies indicate that eosinophils protect cholinergic IMR-32 nerve cells against apoptosis induced by serum deprivation or by the cytokines IL-1β, TNF-α, and IFN-γ. In prior studies, we have shown that prolonged exposure to eosinophils was associated with the long-term survival of both primary cultures of tracheal parasympathetic nerve cells and differentiated IMR-32 cells. That eosinophils were not toxic to nerves was a surprise because these cells contain toxic proteins and release reactive oxygen species, all of which could be toxic to nerve cells (7). To examine the mechanism of eosinophil-mediated neuroprotection, we used eosinophils isolated from peripheral blood and directly cultured them with IMR-32 cells. In addition, to study whether the protection was due to contact between cells or due to factors released from the cells, we prepared eosinophil membranes by removal of nuclear and cytoplasmic material. We have previously shown that these preparations express adhesion molecules and adhere to cholinergic IMR-32 nerve cells in the same manner as viable whole eosinophils (9). These studies showed that eosinophil membranes in coculture with IMR-32 cells prevent neuronal apoptosis, as observed by expression of phosphatidylserine on the outer nerve membrane, caspase-3 activation, and DNA laddering. The protection against apoptosis was prevented by neutralizing Abs to the adhesion molecules ICAM-1 and VCAM-1, alone or in combination. This indicates that protection of IMR-32 cells by eosinophils from apoptosis-inducing agents was mediated by direct cell-cell contact.
Although the studies we conducted established that adhesion of eosinophil membranes to IMR-32 neuronal cells via ICAM-1 and/or VCAM-1 was necessary for the observed protection from apoptosis, we cannot entirely rule out that other eosinophil-specific factors also contribute to nerve cell survival. In contrast to eosinophils and derived membrane preparations, neutrophils did not confer protection from apoptosis on IMR-32 cells. This implies that, for example, cross talk mechanisms between ICAM-1, VCAM-1, and other eosinophil-specific adhesion molecules may be important in adhesion-mediated protection, or that other unidentified eosinophil-specific factors present in the eosinophil membrane preparations are contributing to the observed effects. We established that the membrane preparations used do not contain, for example, the eosinophil degranulation products EPO, EDN, or MBP within the limits of detection of the Abs to which we had access; however, it is possible that other unidentified eosinophil proteins, or degranulation products present at amounts we cannot detect, may play a role in protection.
We have previously shown that ligation of the surface adhesion molecules ICAM-1 and VCAM-1 by eosinophils activates neuronal MAPKs ERK1/2 and p38 with consequent specific activation of the transcription factors NF-κB and AP-1 (10). Using specific inhibitors of these MAPKs, we established that the protective effect was entirely dependent on ERK1/2 and not p38. This is consistent with studies by others, which show that the MEK/ERK pathway plays an important role in cell survival in different cell types (18, 19, 20). We then studied the expression of anti- and proapoptotic genes of the Bcl-2 family, in particular Bfl-1, Bid, and Bad. Bfl-1 is known to be a key regulator of apoptosis in endothelial cells (21, 22) and is thought to mediate its actions by preventing mitochondrial membrane depolarization and sequestering the proapoptotic BH3 domain-only members of the Bcl-2 family (23). The Bfl-1 gene is regulated in neurons and other cell types by NF-κB activation (24, 25, 26), and the MEK/ERK pathway is thought to modulate cell cycle progression and cell survival through its effect on Bfl-1 expression in a number of cell types, including coronary artery smooth muscle and human pancreatic cancer cells (27, 28, 29).
Coincubation of inactivated eosinophil membranes with IMR-32 cells caused up-regulation of the antiapoptotic Bfl-1 gene in the IMR-32 cells. The induction of neuronal Bfl-1 transcription by eosinophils was dependent on adhesion through ICAM-1 and VCAM-1 on the nerve surface and phosphorylation of ERK1/2. Increased expression of Bfl-1 was evident after 2 h of coincubation with eosinophil membranes and was sustained for >24 h (data not shown). Bid and Bad, proapoptotic Bcl-2 family members, were constitutively expressed by IMR-32 cells, but their transcription was not influenced by eosinophil adhesion. These data suggest that eosinophil-mediated protection from apoptosis in the IMR-32 nerves may be due at least in part to up-regulation of the antiapoptotic Bfl-1 gene.
Despite the presence of an inflammatory milieu, there is no evidence of nerve cell death in conditions such as asthma, allergic rhinitis, and inflammatory bowel disease in which eosinophils are found in close association with nerves. In fact, there is evidence of increased vagal reflexes contributing to bronchoconstriction in asthma (30, 31). In addition, eosinophils may be recruited to mucosal nerves in some patients with inflammatory bowel disease in which an increased density of nerve fibers has been observed (32, 33, 34). The current studies show that eosinophils dose dependently protect cholinergic IMR-32 nerve cells against apoptosis induced by serum starvation or exposure to the cytokines that would be seen in inflammatory conditions. The current observations may be part of the mechanism of how peripheral nerves survive in an inflammatory environment.
Although eosinophils have long been implicated in the pathogenesis of a variety of conditions such as asthma and do so by causing epithelial cell dysfunction through the release of granular proteins (7), there is emerging evidence that eosinophils also have a role to play in protection and repair (35, 36). For example, in sensitized animals, they decrease viral replication following a respiratory virus infection (37), and they participate in the recovery from ozone-induced bronchial hyperresponsivness (4). Our observation that cholinergic nerves receive survival signals from eosinophils provides further evidence that the eosinophils have a protective role. The results of the current study suggest that by a process involving recruitment and adhesion to peripheral nerves in the airways and gut, eosinophils may be responsible for the maintenance of cholinergic phenotype in inflamed tissue. This eosinophil-mediated effect on nerve longevity is a further example of how eosinophils may regulate neural plasticity in inflammatory conditions.
Footnotes
-
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
-
↵1 This work was supported by the Health Research Board of Ireland and The Wellcome Trust.
-
↵2 Address correspondence and reprint requests to Dr. Richard W. Costello, Department of Medicine, Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin 9, Ireland. E-mail address: rcostello{at}rcsi.ie
-
↵3 Abbreviations used in this paper: MBP, major basic protein; EDN, eosinophil-derived neurotoxin; EPO, eosinophil peroxidase; PI, propidium iodide.
- Received December 12, 2003.
- Accepted August 26, 2004.
- Copyright © 2004 by The American Association of Immunologists
References
In this issue
