Airway eosinophilia plays a major role in the pathogenesis of asthma with the inhibition of apoptosis by GM-CSF and IL-5 proposed as a mechanism underlying prolonged eosinophil survival. In vivo and ex vivo studies have indicated the capacity of interventions that drive human eosinophil apoptosis to promote the resolution of inflammation. Far less is known about the impact of transendothelial migration on eosinophil survival, in particular, the capacity of endothelial cell-derived factors to contribute toward the apoptosis-resistant phenotype characteristic of airway-resident eosinophils. We examined the effects of conditioned medium from human pulmonary artery endothelial cells (HPAEC-CM) on eosinophil apoptosis in vitro. HPAEC-CM inhibited eosinophil, but not neutrophil apoptosis. This effect was specific to HPAECs and comparable in efficacy to the survival effects of GM-CSF and IL-5. The HPAEC survival factor was shown, on the basis of GM-CSF, IL-5, and IL-3 detection assays, Ab neutralization, and sensitivity to PI3K inhibition, to be clearly discrete from these factors. Gel filtration of HPAEC-CM revealed a peak of eosinophil survival activity at 8–12 kDa, and PCR confirmed the presence of mRNA for CCL5, CCL11, CCL24, CCL26, and CCL27 in the HPAECs. The CCR3 antagonist GW782415 caused a major inhibition of the HPAEC-CM-induced survival effect, and Ab neutralization of individual CCR3 chemokines revealed CCL11 as the major survival factor present in the HPAEC-CM. Furthermore, chemokine Ab arrays demonstrated up-regulation of CCL11 in HPAEC-CM. These data demonstrate the capacity of HPAECs to generate CCR3 agonists and the ability of CCL11 to inhibit human eosinophil apoptosis.

Airway eosinophilia is a characteristic feature of several inflammatory diseases including asthma, allergic rhinitis, nasal polyposis, and eosinophilic bronchitis. In the case of asthma, treatment targeted at eosinophil reduction improves disease control (1, 2, 3). Although many of the mechanisms that result in the selective recruitment, priming, and activation of these cells have been well characterized (4), the processes involved in the elimination of these cells are less well understood. IL-3, IL-5, and GM-CSF are known to play important roles in the maturation, terminal differentiation, activation, and survival of eosinophils, whereas the CCR3 chemokines CCL5 and CCL11 regulate the release of eosinophils from the bone marrow and their migration to sites of allergic inflammation (5, 6, 7).

Potential mechanisms for the removal of eosinophils from tissues include re-entry into the circulation (as recently demonstrated for neutrophils (8)), migration to regional lymph nodes, primary cytolysis, apoptosis, or, in the bowel and airway, transepithelial migration. The importance of the latter mechanism has been demonstrated using allergen challenge in animals, which induces a rapid and highly coordinated exit of eosinophils into the airway lumen estimated at up to 35,000 cells/min/cm2 (9). Once present in the airway lumen, these cells are either expectorated or undergo senescence-related or Fas/Fas ligand-driven apoptosis, which triggers phagocytic removal by alveolar macrophages, inflammatory macrophages, or bronchial epithelial cells (10, 11). Although eosinophils are clearly capable of undergoing constitutive apoptosis both in vitro and in vivo (12, 13, 14), studies examining airway morphology in a variety of allergic inflammatory settings have shown little, if any, evidence of eosinophil apoptosis in the airway wall compartment. Although the highly efficient coupling of granulocyte apoptosis to macrophage removal may result in an underestimation of the true extent of eosinophil apoptosis occurring in vivo, these data support the concept that the capacity of eosinophils to undergo constitutive apoptosis within the airway wall or lung interstitial compartment may be severely restricted, i.e., that these cells adopt an apoptosis-resistant phenotype.

Transmigration of human neutrophils across lung endothelium-epithelium bilayers has been shown to induce a sustained inhibition of apoptosis, which correlates with down-regulation of Fas ligand and TNF-α-R1 expression (15). A similar paradigm has been proposed for the eosinophil because transmigration of these cells across IL-1β-activated human pulmonary microvascular endothelial cells also has a profound impact on eosinophil receptor expression. This process results in the up-regulation of CD69, HLA-DR, and CD54/ICAM-1, together with enhanced respiratory burst activity and prolonged survival (16); this latter effect has been attributed at least in part to the induction and autocrine effects of GM-CSF. In an in vivo setting, the migrating eosinophil is also influenced by a variety of other factors including interactions with other inflammatory and structural cells, cell:matrix contact, and a much broader array of endothelial cell-derived cytokines, colony-stimulating factors, and other proinflammatory molecules including platelet-activating factor, eicosanoids, and NO (17, 18).

To explore the capacity of human pulmonary artery endothelial cells (HPAECs)3 to modulate apoptotic thresholds in human eosinophils, we sought to characterize the effects of HPAEC conditioned medium (HPAEC-CM) in isolated eosinophils in vitro. These studies suggested that unstimulated primary HPAECs secrete a factor or factors distinct from GM-CSF and IL-5 that results in a profound and selective survival advantage for human eosinophils. Further characterization of this survival factor demonstrated the previously unrecognized capacity of HPAECs to secrete the CCR3 ligands CCL5, CCL11, CCL24, CCL26, and CCL27 (RANTES, eotaxin-1, eotaxin-2, eotaxin-3, and cutaneous T cell-attracting chemokine (CTACK)) and for CCL11 to operate as a novel eosinophil survival factor. These data offer new insights into the mechanisms underlying the apoptosis-resistant phenotype of eosinophils within the airway wall and add to the recognized sites of CCR3 active chemokine generation and the functional attributes of CCL11.

Human granulocytes were isolated from the peripheral blood of healthy normal and atopic donors not receiving topical or systemic medication, as previously detailed (19). Approval was obtained from the Cambridge Research Ethics Committee for these studies. Eosinophils were purified from the mixed granulocyte population using immunomagnetic separation with anti-CD16 microbeads (Miltenyi Biotec). Isolated eosinophils were washed in PBS containing Ca2+/Mg2+ and resuspended at 5 × 106 cells/ml in Iscove’s DMEM (Invitrogen Life Technologies) supplemented with 50 U/ml streptomycin and penicillin G and 1× insulin-transferrin-sodium selenite liquid supplement (Sigma-Aldrich).

Granulocytes were harvested at 24 h (unless otherwise stated), cytocentrifuged, fixed in methanol, and stained with Diff-Quik (Greiner). Morphology was examined by light microscopy under oil using a ×100 objective. Apoptotic granulocytes were defined as those with decreased cell size, nuclear and cytoplasmic condensation. A total of 300 cells/slide was counted, with the viewer blinded to the experimental conditions. The number of apoptotic cells was calculated as a percentage of the total cell count. Apoptosis was also assessed by flow cytometry using FITC-labeled Annexin V (BD Biosciences). Stock FITC-Annexin V was diluted 1/100 and propidium iodide (PI) diluted 1/10 with the supplied binding buffer (final concentration, 5 μg/ml). Cells were removed from culture at the times indicated, pelleted by centrifugation (300 × g, 5 min, 4°C), and incubated at 4°C in the dark with 100 μl of the above buffer containing FITC-Annexin V and PI for 15 min. Samples were then diluted with 500 μl of binding buffer and examined using a FACSort (BD Biosciences) by counting 10,000 events per sample, and data were analyzed using FSC Press software (FSC Press) as previously described (20).

HPAECs, HUVECs, human aortic artery endothelial cells (HAECs), and human coronary artery endothelial cells (HCAECs) were obtained from Cambrex. The cells were grown and maintained in endothelial growth medium (EGM-2) supplemented with 2% (v/v) FBS (Invitrogen Life Technologies) and 100 U/ml penicillin G, 100 μg/ml streptomycin, and 25 ng/ml amphotericin B. Cells were used to passage 8 only. Once confluent, cells were washed thoroughly with serum-free Iscove’s DMEM, and then incubated for a further 2 h in serum-free Iscove’s DMEM, followed by a final incubation for 15 h with antibiotic and antimycotic-supplemented serum-free Iscove’s DMEM to generate the conditioned medium (IMDM-CM). Trypan blue was used to confirm that the HPAECs were viable after 15-h serum-free treatment. The CM was removed, centrifuged at 1000 × g (10 min at 4°C), and aliquots were stored at −80°C. Insulin-transferrin-sodium selenite was added as above to all conditioned and control nonconditioned medium before eosinophil or neutrophil incubation. Smooth muscle cells were isolated from the medium of human pulmonary arteries present in lung explants (21) and cultured in DMEM supplemented with 10% (v/v) FBS and the above antibiotics and antimycotics. Human fetal lung fibroblasts (HFL-1) (European Collection of Cell Cultures) were cultured in Ham’s F12 supplemented with 2 mM glutamine, 1× nonessential amino acids, 10% (v/v) FBS, and 100 U/ml penicillin G, 100 μg/ml streptomycin, and 25 ng/ml amphotericin B. All cells were maintained at 37°C in a humidified atmosphere of air containing 5% CO2.

Granulocytes incubated with priming or activating agents under nongradient conditions undergo frustrated chemotaxis, which results in a characteristic change in cell morphology (shape change) and consequent alteration in their forward light scatter properties under flow cytometry (22). Granulocytes were resuspended at 5 × 106 cells/ml in 100 μl of PBS containing Ca2+/Mg2+ and incubated with agonists or buffer for 10 min before fixation in 0.8% (v/v) glutaraldehyde. Percentage shape change was determined by assessing forward scatter under control and activated conditions using FACSort and FSC Press software as detailed previously (23, 24).

HPAECs were grown to confluence in 24-well tissue culture plates and coincubated in the presence or absence of eosinophils (2.5 × 105 per well) resuspended in serum-free Iscove’s DMEM. At 24 h, the eosinophils were removed and apoptosis was assessed by flow cytometry using FITC-labeled Annexin V (see above).

Confirmation of the protein nature of the HPAEC-CM survival factor(s) was sought by determining the effects of both heat treatment (56°C, 45 min) and trypsin digestion. For the latter, HPAEC-CM was incubated with trypsin (1500 ng/ml HPAEC-CM) for 2 h at 37°C and the reaction was terminated by the addition of 1500 ng/ml soya bean trypsin inhibitor before assessing the effect on eosinophil apoptosis.

G-50 Sephadex Superfine beads (Amersham Biosciences) were packed in a 40-cm column (16-mm internal diameter) and equilibrated in phenol red-free Iscove’s DMEM. HPAEC-CM (10 ml) was freeze-dried, stored at −80°C, and resuspended in 550 μl of phenol red-free Iscove’s DMEM. Dextran blue (30 mg/ml) and 50 μl of 125I-labeled cAMP were added to the sample of HPAEC-CM to give a final volume of 0.61 ml. The suspension was spun at 15,000 × g for 5 min at room temperature, and the supernatant was applied to the column. Fractions (1 ml) were collected overnight at 4°C, analyzed for 125I-labeled cAMP and dextran blue markers, and the intervening aliquots were stored at −80°C before coincubation with eosinophils for 24 h.

IL-5 and GM-CSF were measured using an in-house ELISA. Flat-bottom high binding 96-well ELISA plates (Greiner) were coated with 2 μg/ml monoclonal anti-GM-CSF Ab (R&D Systems), diluted in carbonate buffer (0.15 M sodium carbonate, 0.35 M sodium bicarbonate, pH 9.6) for 2 h using a working volume of 50 μl. The plates were washed three times with PBS containing 0.05% Tween 20 (PBS-T) and blocked with 5% heat-inactivated FCS in PBS-T; standards and samples with the relevant controls were then added and incubated overnight at 4°C. Following three washes with PBS-T, biotinylated anti-human GM-CSF (R&D Systems) (0.2 μg/ml in PBS-T plus 5% FCS) was added, and the incubations continued for 2 h at room temperature. The plates were washed three times and ExtraAvidin alkaline phosphatase conjugate (1:250) (Sigma-Aldrich) added and incubated at room temperature for 2 h. After two further washes, 1 mg/ml p-nitrophenyl phosphate (Sigma-Aldrich) was added diluted in diethanolamine buffer. The latter buffer was prepared by dissolving 101 mg of MgCl2·6H2O in 800 ml of distilled H2O before the addition of 97 ml of diethanolamine and adjustment of the pH to 9.8. Plates were read at 405 nm using a Bio-Rad 550 microplate reader and the data were analyzed using MPM III Vs 1.57 software. The detection limit of this ELISA for GM-CSF was 30 pg/ml. The IL-5 ELISA was performed as described above with the substitution of IL-5 mAb at a coating concentration of 2 μg/ml and the use of monoclonal anti-goat/sheep-alkaline phosphatase (GT-34) (Sigma-Aldrich) (1:400). The detection limit of this assay for IL-5 was 100 pg/ml.

The presence of chemokines in HPAEC-CM and HUVEC-CM was assessed using the human protein chemokine array kit (RayBiotech) (see Fig. 5) as detailed in the manufacturer’s instructions. The resultant blots were quantified by densitometry (Scion Image; Scion Corporation). The levels of CCL11 were also measured using a CCL11 Duoset ELISA (R&D Systems) according to the manufacturer’s instructions. The HPAEC-CM and HUVEC-CM were also analyzed by Rules-Based Medicine using the Luminex FlowMetrix system.

FIGURE 5.

HPAEC-CM-derived eosinophil survival factor is a CCR3 agonist. A, Representative chemokine membrane array film showing HPAEC-CM. Grid references A1, B1, and L4 represent the internal positive controls, and the circled spots represent CCL11 (K1) and IL-8 (I2). B, Densitometry analysis of chemokine arrays with grid references. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with HUVEC-CM values). The mean values were calculated as a percentage of the positive controls (A1 and B1). C, Effect of GW782415 on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), GW782415 (10 μM), or with GM-CSF plus GW782415 (10 μM) were incubated for 30 min at 37°C before resuspension with eosinophils. HPAEC-CM alone, with GW782415 (10 μM), or with its diluent (10 μM DMSO) was also incubated for 30 min at 37°C before resuspension with eosinophils. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of four independent experiments (∗, p < 0.01 compared with HPAEC-CM values). D, The effect of M3 on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM (control), GM-CSF (10 ng/ml), M3 (1 μg/ml), CrmE (1 μg/ml), GM-CSF plus M3 (1 μg/ml), or with GM-CSF plus CrmE (1 μg/ml) were incubated for 1 h at 37°C before resuspension with eosinophils. HPAEC-CM alone, with M3 (1 μg/ml), or with CrmE (1 μg/ml) was also incubated for 2 h at room temperature before resuspension with eosinophils. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SD of two independent experiments. Limited availability of M3 precluded additional experimental repeats.

FIGURE 5.

HPAEC-CM-derived eosinophil survival factor is a CCR3 agonist. A, Representative chemokine membrane array film showing HPAEC-CM. Grid references A1, B1, and L4 represent the internal positive controls, and the circled spots represent CCL11 (K1) and IL-8 (I2). B, Densitometry analysis of chemokine arrays with grid references. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with HUVEC-CM values). The mean values were calculated as a percentage of the positive controls (A1 and B1). C, Effect of GW782415 on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), GW782415 (10 μM), or with GM-CSF plus GW782415 (10 μM) were incubated for 30 min at 37°C before resuspension with eosinophils. HPAEC-CM alone, with GW782415 (10 μM), or with its diluent (10 μM DMSO) was also incubated for 30 min at 37°C before resuspension with eosinophils. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of four independent experiments (∗, p < 0.01 compared with HPAEC-CM values). D, The effect of M3 on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM (control), GM-CSF (10 ng/ml), M3 (1 μg/ml), CrmE (1 μg/ml), GM-CSF plus M3 (1 μg/ml), or with GM-CSF plus CrmE (1 μg/ml) were incubated for 1 h at 37°C before resuspension with eosinophils. HPAEC-CM alone, with M3 (1 μg/ml), or with CrmE (1 μg/ml) was also incubated for 2 h at room temperature before resuspension with eosinophils. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SD of two independent experiments. Limited availability of M3 precluded additional experimental repeats.

Close modal

In the cytokine and chemokine neutralization experiments, all Abs and isotype-matched IgG controls were obtained from R&D Systems; HPAEC-CM was preincubated with Ab at the concentration specified for 1 h at 37°C before resuspension with eosinophils (5 × 106/ml) and further culture. The following Ab and IgG concentrations were used: IL-5 and GM-CSF at 50 μg/ml; CCL27, CCL11, CCL24, CCL26, and CCL5 at 10 μg/ml. For the PI3K inhibition experiments 1 μM LY294002 (Calbiochem) was preincubated with the eosinophils for 15 min before the addition of medium. The pan chemokine blocker M3 and control viral TNF-α receptor cytokine response modifier E (CrmE) were donated by Dr. A. Alcami (University of Cambridge, Cambridge, U.K.). The CrmE protein was generated and purified using the same conditions as M3. GW782415 was a gift from Dr. C. Horgan (GlaxoSmithKline, Stevenage, U.K.).

Total RNA was isolated from the HPAECs using RNeasy columns (Qiagen) with the additional on column DNA digest step (RNase-Free DNase Set; Qiagen). RNA (in micrograms) was transcribed into cDNA using a first-strand cDNA synthesis kit (AccuScript; Stratagene). IQ SYBR-green supermix (Bio-Rad Laboratories) was used with primers for CCL5, CCL11, CCL24, CCL26, and CCL27 (Qiagen) for quantitative PCR (iCycler IQ system; Bio-Rad Laboratories). Control primers for β-actin were designed in-house: sense, GCACCACACCTTCTACAATGA, antisense, TGTCACGCACGATTTCCC; product length, 400 bp. The mean cycle thresholds (Ct) were determined for the genes of interest and endogenous β-actin.

Unless otherwise stated, all data are expressed as mean ± SEM of (n) independent experiments, each performed in duplicate or triplicate. Data were analyzed using GraphPad Prism and Student’s t test. A value of p < 0.05 was considered significant.

As previously described (6, 25), coincubation of human eosinophils for 24 h with either GM-CSF or IL-5 results in a substantial inhibition of apoptosis (Fig. 1,A); this effect was evident using either direct morphological quantification (Fig. 1, A and B) or Annexin V staining (Fig. 1,C). Culture of eosinophils for 24 h under otherwise identical conditions in medium that had been in contact with a monolayer of HPAECs for 15 h also inhibited eosinophil apoptosis (Fig. 1). The efficacy of the HPAEC-CM survival of this effect was comparable with that of GM-CSF and IL-5. Of note, all control, IL-5, and GM-CSF incubations were performed using medium that had been handled in an identical manner aside from exposure to HPAECs.

FIGURE 1.

Conditioned medium from HPAECs contains an eosinophil survival factor. A, Effect of HPAEC-CM on eosinophil apoptosis. Eosinophils (5 × 106/ml) were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), GM-CSF (10 ng/ml), or with HPAEC-CM. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). B, Morphology of eosinophils cultured in HPAEC-CM. Photomicrographs (original magnification, ×400 magnification) of cytospin preparations. Eosinophils were incubated in Iscove’s DMEM alone, HPAEC-CM, or unconditioned medium containing IL-5 (10 ng/ml) or GM-CSF (10 ng/ml). Eosinophils were harvested at 24 h. Examples of apoptotic eosinophils are indicated with closed arrows. C, Effect of HPAEC-CM on eosinophil apoptosis (FACS analysis). Eosinophils were incubated in Iscove’s DMEM alone, HPAEC-CM, or unconditioned medium containing IL-5 (10 ng/ml) or GM-CSF (10 ng/ml). Cells were harvested at 24 h, and apoptosis was assessed by flow cytometry using FITC-labeled human Annexin V and PI. Histograms showing Annexin V binding are from one experiment representative of two others.

FIGURE 1.

Conditioned medium from HPAECs contains an eosinophil survival factor. A, Effect of HPAEC-CM on eosinophil apoptosis. Eosinophils (5 × 106/ml) were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), GM-CSF (10 ng/ml), or with HPAEC-CM. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). B, Morphology of eosinophils cultured in HPAEC-CM. Photomicrographs (original magnification, ×400 magnification) of cytospin preparations. Eosinophils were incubated in Iscove’s DMEM alone, HPAEC-CM, or unconditioned medium containing IL-5 (10 ng/ml) or GM-CSF (10 ng/ml). Eosinophils were harvested at 24 h. Examples of apoptotic eosinophils are indicated with closed arrows. C, Effect of HPAEC-CM on eosinophil apoptosis (FACS analysis). Eosinophils were incubated in Iscove’s DMEM alone, HPAEC-CM, or unconditioned medium containing IL-5 (10 ng/ml) or GM-CSF (10 ng/ml). Cells were harvested at 24 h, and apoptosis was assessed by flow cytometry using FITC-labeled human Annexin V and PI. Histograms showing Annexin V binding are from one experiment representative of two others.

Close modal

To determine whether the atopy status of the eosinophil donor could alter the HPAEC-CM-induced eosinophil survival response, data from 10 atopic and 10 nonatopic donors were analyzed separately. There were no significant differences between the two groups with regards to basal apoptosis rates or the survival effect induced by HPAEC-CM (data not shown).

To investigate the specificity of the above eosinophil survival to HPAECs, conditioned medium was prepared in an identical manner from a variety of other nonpulmonary human endothelial cells including HUVECs, aorta (HAECs), and coronary artery (HCAECs). As shown in Fig. 2, A and B, these conditioned media had no effect on eosinophil apoptosis when assessed at 24 h. Conditioned medium from human lung fibroblasts (HFL-1) and smooth muscle cells derived from human pulmonary arteries likewise had no effect on eosinophil survival (data not shown). In contrast, HPAECs, HUVECs, and HCAECs all inhibited eosinophil apoptosis when the two cell types were in direct contact (Fig. 2 C). Hence, whereas endothelial cells appear to display a generic capacity to prolong eosinophil survival when cultured together, HPAECs possess an additional and distinct survival mechanism mediated by a soluble factor.

FIGURE 2.

Specificity of HPAEC-CM-induced eosinophil survival effect. A, Effect of HUVEC-CM on eosinophil apoptosis. Eosinophils were incubated with IL-5 (10 ng/ml), HPAEC-CM, and HUVEC-CM. Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with control values). B, Effect of HCAEC-CM and HAEC-CM on eosinophil apoptosis. Eosinophils were incubated in HAEC-CM or HCAEC-CM. Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05 compared with control values). C, Effect of eosinophil-endothelial cell contact on eosinophil apoptosis. HPAECs, HUVECs, HAECs, and HCAECs were grown to confluence, and the medium was replaced with serum-free Iscove’s DMEM with or without eosinophils (2.5 × 105 eosinophils per well). Eosinophils were harvested at 24 h, and apoptosis was assessed by FACS analysis. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with control values). D, Effect of HPAEC-CM on constitutive neutrophil apoptosis. Neutrophils (5 × 106/ml) were incubated in Iscove’s DMEM alone, with GM-CSF (10 ng/ml) or HPAEC-CM. Neutrophils were harvested at 6 h (□) and at 20 h (▪). Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). E, Effect of HPAEC-CM on neutrophil shape change. Neutrophils (5 × 106/ml) were incubated for 10 min, and shape change was assessed as described in Materials and Methods. i, Histogram showing the effect of PBS alone or fMLP (1 μM) on neutrophil shape change from one experiment (representative of two others). ii, Effect of PBS, fMLP (1 μM), HPAEC-CM, fMLP plus HPAEC-CM, IMDM (Iscove’s DMEM), and fMLP plus IMDM on neutrophil shape change. Percentage shape change was calculated based on the percentage of cells in the M2 gate. Data represent the mean ± SEM of three independent experiments, each performed in triplicate.

FIGURE 2.

Specificity of HPAEC-CM-induced eosinophil survival effect. A, Effect of HUVEC-CM on eosinophil apoptosis. Eosinophils were incubated with IL-5 (10 ng/ml), HPAEC-CM, and HUVEC-CM. Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with control values). B, Effect of HCAEC-CM and HAEC-CM on eosinophil apoptosis. Eosinophils were incubated in HAEC-CM or HCAEC-CM. Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05 compared with control values). C, Effect of eosinophil-endothelial cell contact on eosinophil apoptosis. HPAECs, HUVECs, HAECs, and HCAECs were grown to confluence, and the medium was replaced with serum-free Iscove’s DMEM with or without eosinophils (2.5 × 105 eosinophils per well). Eosinophils were harvested at 24 h, and apoptosis was assessed by FACS analysis. Data represent the mean ± SEM of three independent experiments (∗, p < 0.05 compared with control values). D, Effect of HPAEC-CM on constitutive neutrophil apoptosis. Neutrophils (5 × 106/ml) were incubated in Iscove’s DMEM alone, with GM-CSF (10 ng/ml) or HPAEC-CM. Neutrophils were harvested at 6 h (□) and at 20 h (▪). Apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). E, Effect of HPAEC-CM on neutrophil shape change. Neutrophils (5 × 106/ml) were incubated for 10 min, and shape change was assessed as described in Materials and Methods. i, Histogram showing the effect of PBS alone or fMLP (1 μM) on neutrophil shape change from one experiment (representative of two others). ii, Effect of PBS, fMLP (1 μM), HPAEC-CM, fMLP plus HPAEC-CM, IMDM (Iscove’s DMEM), and fMLP plus IMDM on neutrophil shape change. Percentage shape change was calculated based on the percentage of cells in the M2 gate. Data represent the mean ± SEM of three independent experiments, each performed in triplicate.

Close modal

Of particular interest, HPAEC-CM had no effect on the rate of neutrophil apoptosis assessed at either 6 or 20 h (Fig. 2,D). Likewise, there was no effect of the HPAEC-CM on neutrophil shape change assessed by flow cytometry (Fig. 2 E, i and ii) or by light microscopy (data not shown), an assay that is highly sensitive in detecting agents that cause neutrophil priming or activation.

HPAEC-CM concentration-response experiments identified loss of the eosinophil survival effect at HPAEC-CM dilutions of 1/10 dilution or greater (Fig. 3,A). This implies that despite possessing an antiapoptotic efficacy comparable with GM-CSF and IL-5, this survival factor(s) displays either low potency or is expressed in low abundance by the HPAEC-CM. Time course experiments shown in Fig. 3,B revealed that a minimum conditioning period of 15 h was required to demonstrate the HPAEC-CM eosinophil survival. Hence, the HPAEC-CM-mediated eosinophil survival effect was both concentration and time dependent. Likewise, heat inactivation of HPAEC-CM or pretreatment with trypsin completely abrogated the survival effect (Fig. 3, C and D), indicating the likely protein nature of the HPAEC-derived survival factor.

FIGURE 3.

Time-dependent generation and protein nature of HPAEC-CM survival effect. A, Effect of different dilutions of HPAEC-CM on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), HPAEC-CM (1/1), or HPAEC-CM diluted 1/3, 1/10, or 1/30 with Iscove’s DMEM. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05; ∗∗, p < 0.01 compared with control values). B, Effect of HPAEC culture period on the HPAEC-CM-induced survival effect. Eosinophils were incubated with DMEM alone (control), or HPAEC-CM removed following 3, 6, 9, or 15 h of contact with HPAECs. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05 compared with control values). C, The effect of trypsin digestion on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM with soya bean trypsin inhibitor (SBTI) alone or with trypsin plus SBTI, HPAEC-CM with SBTI alone or with trypsin plus SBTI were digested as described in Materials and Methods. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM from three separate experiments (∗, p < 0.05 compared with HPAEC-CM values). D, The effect of heat inactivation on HPAEC-CM-induced eosinophil survival. Eosinophils were incubated in Iscove’s DMEM alone (control), HPAEC-CM, or their heat-inactivated equivalents. DMEM and HPAEC-CM were inactivated by heating at 56°C for 45 min. Data represent the mean ± SEM from three independent experiments (∗, p < 0.05 compared with control values).

FIGURE 3.

Time-dependent generation and protein nature of HPAEC-CM survival effect. A, Effect of different dilutions of HPAEC-CM on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), HPAEC-CM (1/1), or HPAEC-CM diluted 1/3, 1/10, or 1/30 with Iscove’s DMEM. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05; ∗∗, p < 0.01 compared with control values). B, Effect of HPAEC culture period on the HPAEC-CM-induced survival effect. Eosinophils were incubated with DMEM alone (control), or HPAEC-CM removed following 3, 6, 9, or 15 h of contact with HPAECs. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in duplicate (∗, p < 0.05 compared with control values). C, The effect of trypsin digestion on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM with soya bean trypsin inhibitor (SBTI) alone or with trypsin plus SBTI, HPAEC-CM with SBTI alone or with trypsin plus SBTI were digested as described in Materials and Methods. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM from three separate experiments (∗, p < 0.05 compared with HPAEC-CM values). D, The effect of heat inactivation on HPAEC-CM-induced eosinophil survival. Eosinophils were incubated in Iscove’s DMEM alone (control), HPAEC-CM, or their heat-inactivated equivalents. DMEM and HPAEC-CM were inactivated by heating at 56°C for 45 min. Data represent the mean ± SEM from three independent experiments (∗, p < 0.05 compared with control values).

Close modal

GM-CSF and IL-5 are both potent eosinophil survival cytokines expressed by HPAECs and eosinophils (26, 27). Consequently, IL-5 and GM-CSF were considered to be the most likely candidates for the HPAEC-derived eosinophil survival factor. However, GM-CSF and IL-5 levels, when measured by ELISA in three separately prepared CMs, did not differ between the HPAEC-CM and HUVEC-CM (which had no survival influence on human eosinophils) (Fig. 4 A). Furthermore, for GM-CSF in particular, the levels were below the concentration reported elsewhere to induce eosinophil survival (28). Independent analysis of HPAEC-CM and HUVEC-CM samples by multiplexed particle-based flow cytometry again revealed no difference in the levels of these two cytokines in the two conditioned medium (GM-CSF: HPAEC-CM, 27 pg/ml; HUVEC-CM, 42 pg/ml; IL-5: HPAEC-CM, 31 pg/ml; HUVEC-CM, 28 pg/ml). IL-3, a further well-characterized eosinophil survival cytokine (28), was only detected by multiplexed flow cytometry at sub-picogram per milliliter amounts and did not differ between HPAEC-CM and HUVEC-CM (0.3 and 0.6 pg/ml, respectively). In addition, incorporation of lactose into our assay system, which blocks the eosinophil survival affects of the galectin-like molecule ecalectin (29), had no influence on the HPAEC-CM survival effect (data not shown).

FIGURE 4.

HPAEC-CM-induced eosinophil survival effect is not mediated by IL-5 or GM-CSF. A, ELISA measurement of GM-CSF and IL-5 in HPAEC-CM. As a positive control, PHA-stimulated PBMCs were used. Data represent the mean ± SEM of three independent experiments. B, The effect of GM-CSF and IL-5 neutralizing Abs on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), anti-GM-CSF (50 μg/ml), GM-CSF (10 ng/ml) plus anti-GM-CSF (50 μg/ml), or with IgG (50 μg/ml) were incubated for 1 h at 37°C before resuspension with eosinophils. HPAEC-CM alone, with anti-GM-CSF (50 μg/ml), or with IgG (50 μg/ml) was also incubated for 1 h at 37°C before resuspension with eosinophils. Anti-IL-5 was also used as described for GM-CSF and at 50 μg/ml. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SD of two independent experiments. C, Effect of LY294002 on HPAEC-CM-induced eosinophil survival. Eosinophils were preincubated with or without LY294002 (1 μM) for 15 min before resuspension in Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), HPAEC-CM alone, LY294002, GM-CSF plus LY294002, or HPAEC-CM plus LY294002. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values).

FIGURE 4.

HPAEC-CM-induced eosinophil survival effect is not mediated by IL-5 or GM-CSF. A, ELISA measurement of GM-CSF and IL-5 in HPAEC-CM. As a positive control, PHA-stimulated PBMCs were used. Data represent the mean ± SEM of three independent experiments. B, The effect of GM-CSF and IL-5 neutralizing Abs on HPAEC-CM-induced eosinophil survival. Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), anti-GM-CSF (50 μg/ml), GM-CSF (10 ng/ml) plus anti-GM-CSF (50 μg/ml), or with IgG (50 μg/ml) were incubated for 1 h at 37°C before resuspension with eosinophils. HPAEC-CM alone, with anti-GM-CSF (50 μg/ml), or with IgG (50 μg/ml) was also incubated for 1 h at 37°C before resuspension with eosinophils. Anti-IL-5 was also used as described for GM-CSF and at 50 μg/ml. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SD of two independent experiments. C, Effect of LY294002 on HPAEC-CM-induced eosinophil survival. Eosinophils were preincubated with or without LY294002 (1 μM) for 15 min before resuspension in Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), HPAEC-CM alone, LY294002, GM-CSF plus LY294002, or HPAEC-CM plus LY294002. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values).

Close modal

Although the concentrations of IL-5 and GM-CSF did not differ between the HPAEC-CM and HUVEC-CM, both were detectable; likewise, IL-5 has been reported to inhibit eosinophil apoptosis at concentrations as low as 10 pg/ml (30). For this reason, cytokine Ab neutralization experiments were performed. As shown in Fig. 4,B, whereas anti-IL-5 and anti-GM-CSF completely blocked the eosinophil survival effect of 10 ng/ml IL-5 and 10 ng/ml GM-CSF, respectively, neither Ab was able to influence the HPAEC-CM survival effect. Furthermore, coincubation of eosinophils with the phosphoinositide 3-kinase inhibitor LY294002 (1 μM) prevented GM-CSF-mediated inhibition of eosinophil apoptosis but failed to modulate the HPAEC-CM-induced eosinophil survival response (Fig. 4 C). Together, these data suggest that the HPAEC-CM survival response is not mediated by GM-CSF, IL-5, or IL-3.

Gel filtration chromatography followed by assay of the individual fractions for inhibition of eosinophil survival demonstrated a molecular mass for the HPAEC-CM survival factor of ∼8–12 kDa (data not shown), suggesting the involvement of a chemokine. Three complementary strategies were then pursued: 1) use of a human chemokine Ab array assay to detail the chemokines present in HPAEC-CM and HUVEC-CM, 2) examination of the effects of chemokine inhibition on the HPAEC-CM survival effect using the herpes virus-encoded broad spectrum secreted chemokine binding protein M3 (31), and 3) examination of the effects of the CCR3 antagonist GW782415 on the HPAEC-CM survival effect.

Using the human chemokine Ab array assay, IL-8 and CCL11 (eotaxin-1) were identified as the only chemokines that were present to a greater extent in the HPAEC-CM compared with HUVEC-CM samples (Fig. 5, A and B). Importantly, this assay revealed no differences in the levels of the CCR3 chemokines CCL5, CCL24, CCL26, CCL27, MCP-2, -3, or -4 (Fig. 5 B). Given that eosinophils do not express receptors for IL-8 (32), a direct role for this cytokine in the HPAEC-CM survival effect was considered unlikely.

The ability of GW782415 to block specifically CCR3-mediated eosinophil responses has been established in tests by GlaxoSmithKline and by Fryer et al. (33). Furthermore, the ability of GW782415 to block CCR3 responses was confirmed by examining eosinophil shape change responses in whole blood, where this compound fully inhibited CCL11-induced eosinophil polarization (data not shown). As shown in Fig. 5 C, GW782415 (10 μM) caused a ∼70% reduction of the HPAEC-CM-mediated inhibition of eosinophil apoptosis without affecting the rate of apoptosis under basal or GM-CSF-stimulated conditions.

The broad-spectrum chemokine inhibitor M3 (31) was likewise used to investigate the role of chemokines in the eosinophil-survival response. Although M3 was able to fully block IL-8-mediated superoxide anion production and enhanced survival in neutrophils (data not shown), M3 had no effect on the survival effect of HPAEC-CM (or GM-CSF) in eosinophils (Fig. 5,D). Of note, however, coincubation of purified eosinophils with CrmE, a control poxvirus-derived soluble TNFR (34) prepared under identical conditions to M3, imparted a significant survival effect (Fig. 5 D), and although M3 binds to most of the C, CC, CXC, and CX3C family proteins, the individual binding affinities vary and for eotaxin-1 is >1000 pM (Dr. A. Alcami, unpublished observations).

Quantitative PCR demonstrated that HPAECs express mRNA for CCL5, CCL11, CCL24, CCL26, and CCL27 (Fig. 6,A), supporting the possibility that one or more of these CCR3 agonists might underlie the HPAEC-CM eosinophil survival effect. To verify that CCR3 agonists have the capacity to influence apoptotic thresholds in vitro, we first used coculture experiments using CCL5, CCL11, CCL24, CCL26, and CCL27 at concentrations reported to be optimal for previously described biological actions. In contrast to previous reports (35), CCL5, CCL11, CCL24, CCL26, and CCL27 all inhibited eosinophil apoptosis at 24 h, although to a lesser extent than GM-CSF (Fig. 6,B). To confirm that CCL11, which was the only CCR3 agonist differentially expressed between the HPAEC-CM and the HUVEC-CM, was responsible for the HPAEC-CM eosinophil survival effect, individual Ab neutralization studies were performed (Table I and Fig. 6,C). This demonstrated that anti-CCL11 Abs but not Abs against CCL5, CCL24, CCL26, or CCL27 could fully reverse the HPAEC-CM-mediated eosinophil survival effect. Furthermore, the level of CCL11 detected in the HPAEC-CM by ELISA was 40 pg/ml (Fig. 7,A). The addition of 40 pg/ml CCL11 to HUVEC-CM resulted in an eosinophil survival response comparable with that conferred by HPAEC-CM (Fig. 7 B). These results highlight a potential role for HPAEC-derived CCL11 as an in vivo modulator of eosinophil longevity.

FIGURE 6.

Identification of CCL11 as a novel HPAEC-derived product capable of causing survival of human eosinophils. A, The mRNA expression of CCR3 agonists by HPAECs. Quantitative PCR cycle thresholds for CCL5, CTACK, CCL11, CCL24, and CCL26 mRNA. β-Actin mRNA expression was detected as a control. Confluent HPAECs were serum-starved for 15 h as described in Materials and Methods before RNA isolation. Data represent the mean of duplicate samples from a single experiment with the SD values all under 5% of the mean value. B, The effect of CCR3 agonists on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), CCL5 (200 ng/ml), CCL27 (10 μg/ml), CCL11 (200 ng/ml), CCL24 (500 ng/ml), and CCL26 (5 μg/ml). Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of seven independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). Values represent mean ± SEM apoptosis values expressed as a percentage of the control value. C, The effect of CCL11 Ab neutralization on the HPAEC-CM-induced eosinophil survival effect. Samples were preincubated with the Ab or isotype-matched IgG controls for 1 h at 37°C before resuspension with eosinophils (5 × 106/ml) as described in Materials and Methods. Values represent mean ± SEM apoptosis values expressed as a percentage of the control value (∗, p < 0.05 compared with HPAEC-CM values).

FIGURE 6.

Identification of CCL11 as a novel HPAEC-derived product capable of causing survival of human eosinophils. A, The mRNA expression of CCR3 agonists by HPAECs. Quantitative PCR cycle thresholds for CCL5, CTACK, CCL11, CCL24, and CCL26 mRNA. β-Actin mRNA expression was detected as a control. Confluent HPAECs were serum-starved for 15 h as described in Materials and Methods before RNA isolation. Data represent the mean of duplicate samples from a single experiment with the SD values all under 5% of the mean value. B, The effect of CCR3 agonists on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), GM-CSF (10 ng/ml), CCL5 (200 ng/ml), CCL27 (10 μg/ml), CCL11 (200 ng/ml), CCL24 (500 ng/ml), and CCL26 (5 μg/ml). Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of seven independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values). Values represent mean ± SEM apoptosis values expressed as a percentage of the control value. C, The effect of CCL11 Ab neutralization on the HPAEC-CM-induced eosinophil survival effect. Samples were preincubated with the Ab or isotype-matched IgG controls for 1 h at 37°C before resuspension with eosinophils (5 × 106/ml) as described in Materials and Methods. Values represent mean ± SEM apoptosis values expressed as a percentage of the control value (∗, p < 0.05 compared with HPAEC-CM values).

Close modal
Table I.

The effect of CCR3 chemokine neutralizing Abs on HPAEC-CM-induced eosinophil survivala

No AbPlus Neutralizing Ab
Ab DetailsControlGM-CSFHPAEC-CMIgGHPAEC-CM + IgGAb aloneHPAEC-CMGM-CSF
Anti-CCL11 100 38 ± 4 33 ± 3 95 ± 23 41 ± 1NS 100 ± 16 105 ± 12b 38 ± 4 
Anti-CCL24 100 39 ± 5 41 ± 2 75 ± 13 59 ± 6NS 98 ± 4 72 ± 20NS 40 ± 15 
Anti-CCL26 100 37 ± 1 58 ± 9 109 ± 24 65 ± 5NS 105 ± 10 89 ± 15NS 62 ± 27 
Anti-CCL5 100 35 ± 0 50 ± 8 84 ± 12 65 ± 2NS 100 ± 9 87 ± 10NS 53 ± 8 
Anti-CCL27 100 28 ± 2 39 ± 4 76 ± 0 44 ± 3NS 69 ± 9.5 42 ± 2NS 41 ± 5 
No AbPlus Neutralizing Ab
Ab DetailsControlGM-CSFHPAEC-CMIgGHPAEC-CM + IgGAb aloneHPAEC-CMGM-CSF
Anti-CCL11 100 38 ± 4 33 ± 3 95 ± 23 41 ± 1NS 100 ± 16 105 ± 12b 38 ± 4 
Anti-CCL24 100 39 ± 5 41 ± 2 75 ± 13 59 ± 6NS 98 ± 4 72 ± 20NS 40 ± 15 
Anti-CCL26 100 37 ± 1 58 ± 9 109 ± 24 65 ± 5NS 105 ± 10 89 ± 15NS 62 ± 27 
Anti-CCL5 100 35 ± 0 50 ± 8 84 ± 12 65 ± 2NS 100 ± 9 87 ± 10NS 53 ± 8 
Anti-CCL27 100 28 ± 2 39 ± 4 76 ± 0 44 ± 3NS 69 ± 9.5 42 ± 2NS 41 ± 5 
a

Cell medium was preincubated with the Ab or isotype-matched IgG control for 1 h at 37° C before resuspension with eosinophils as described in Materials and Methods. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Values represent apoptosis expressed as a percentage of the control value and are the mean ± SEM of three independent experiments (

b

, p < 0.05 compared with HPAEC-CM values). Note that each set of neutralization experiments was conducted separately due to limitations on human eosinophil number.

FIGURE 7.

The effect of CCL11 addition to HUVEC-CM in mediating eosinophil apoptosis. A, ELISA measurement of CCL11 in HPAEC-CM. As a positive control, TNF-α-stimulated PBMCs were used. Data represent the mean ± SEM of three independent experiments each performed in duplicate. B, Effect of the addition of CCL11 to HUVEC-CM on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), HPAEC-CM, HUVEC-CM, or HUVEC-CM containing 40, 400, 4000 pg/ml, or 200 ng/ml CCL11. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values).

FIGURE 7.

The effect of CCL11 addition to HUVEC-CM in mediating eosinophil apoptosis. A, ELISA measurement of CCL11 in HPAEC-CM. As a positive control, TNF-α-stimulated PBMCs were used. Data represent the mean ± SEM of three independent experiments each performed in duplicate. B, Effect of the addition of CCL11 to HUVEC-CM on eosinophil apoptosis. Eosinophils were incubated in Iscove’s DMEM alone (control), IL-5 (10 ng/ml), HPAEC-CM, HUVEC-CM, or HUVEC-CM containing 40, 400, 4000 pg/ml, or 200 ng/ml CCL11. Eosinophils were harvested at 24 h, and apoptosis was assessed morphologically. Data represent the mean ± SEM of three independent experiments, each performed in triplicate (∗, p < 0.05 compared with control values).

Close modal

This study demonstrates the capacity of HPAECs to secrete an array of chemokines that activate CCR3, including CCL5, CCL11, CCL24, CCL26, and CCL27, and for these CCR3 ligands to delay human eosinophil apoptosis in vitro.

This observation informs directly the current debate regarding the capacity of eosinophils to undergo apoptosis in human airway tissues. Hence, Persson, Erjefalt, and colleagues (36, 37) have provided compelling evidence that, although eosinophil cytolysis is observed in the airway mucosa, most eosinophils leave via transepithelial migration and that apoptosis is only detected in cells present in the airway lumen. This observation also holds true following corticosteroid treatment, which is a powerful inducer of eosinophil apoptosis both in vitro and within the airway lumen (2, 12). These findings, together with the recognized differences in the morphology of circulating and tissue eosinophils, have suggested that the chemoattraction, transendothelial migration, and tissue residency of eosinophils may cause a significant change in their capacity to undergo apoptosis (16). Previous studies have suggested that this phenotypic switch may relate to the exposure of migrating cells to GM-CSF (16), IL-3 (38), IL-5 (19), IL-13 (39), IL-15 (40), or leukotriene B4 (41), and occur through changes in the phosphorylation of Syk and Lyn and activation of the p21 Ras/Raf-1, Jak/Stat, PI3K, and NF-κB pathways (13, 42, 43, 44). The concept that the microenvironment present within the inflamed airway wall may confer a primed and apoptosis-resistant phenotype in the eosinophil (45) has also been proposed as a potential mechanism responsible for the persistence of tissue eosinophilia following anti-IL-5 and IL-12 treatment in asthmatic subjects (46, 47).

The above studies initially suggested that GM-CSF might account for the observed survival effect of HPAEC-CM in human eosinophils; this possibility was excluded, however, by the presence of equivalent and low picogram per milliliter amounts of GM-CSF in both HPAEC-CM and HUVEC-CM and, more convincingly, by the failure of anti-GM-CSF neutralizing Abs or the PI3K inhibitor, LY294002, to modulate the HPAEC-CM survival response. Moreover, a role for GM-CSF would not have supported the very clear eosinophil over neutrophil selectivity observed in the survival effect. A similar set of data was obtained to exclude the involvement of IL-5. Knowledge of the protein nature, predicted size (8–12 kDa), and the major blockade of the biological effect by the selective CCR3 antagonist GW782451 focused our attention on the CC family of chemokines and led to the detection of CCL11, CCL24, CCL26, CTACK, and RANTES in the HPAEC-CM. Although all five of these chemokines were shown to display the capacity to induce eosinophil survival in vitro, only neutralization of CCL11 caused complete loss of the survival efficacy of the HPAEC-CM.

Our data appear to contradict earlier reports that eotaxin-1 enhances survival only in mouse eosinophils (29, 48, 49). In the latter study, which was conducted in Ag-exposed BALB/c mice, the spontaneous survival of eosinophils purified from the bronchoalveolar lavage fluid was enhanced in the presence of the anti-CCR3 Ab 23321A (an effect not observed with CCR3 antagonist GW782451 in the current study), and the authors demonstrated a concentration-dependent survival effect of rhCCL11, CCL24, and CCL26, most marked when CCL11 and CCL26 were applied together. However, the magnitude of the survival effect was significantly lower than that induced by GM-CSF and only observed at concentrations of CCL11 at 150 ng/ml; moreover, the effect was also ablated by anti-GM-CSF Abs suggesting an autocrine effect again quite unlike that observed in the current study. We recognize that the eosinophil survival effect of exogenous CCL11 (eotaxin-1) alone was not as efficacious as GM-CSF; this may reflect an additional more modest contribution from the other CCR3 agonists present within the HPAEC-CM. The reasons underlying the lack of effect of rhCCL11 on eosinophil survival in previous studies using human peripheral blood-derived cells is uncertain but may relate to differences in cell purification and/or culture techniques.

This study therefore adds to the repertoire of biological effects of CCL11 and points to a new and important site of production, namely the pulmonary artery endothelium. CCL11 (eotaxin-1) is recognized to mediate eosinophil chemotaxis and activation in both mice and humans and is thought to play a central role in eosinophil recruitment to the airways of Ag-challenged animals (50, 51). CCL11 is also recognized to induce chemotaxis in basophils, Th2 lymphocytes, airway smooth muscle cells, and mast cells (52, 53). Likewise, in humans, CCL11 can be readily detected in the sputum of patients with moderate and severe asthma and accounts for ∼50% of the total eosinophil chemotactic activity present in such samples (54). It is perhaps not surprising therefore that we now reveal an additional antiapoptotic capacity of this important human chemokine.

Previously identified sites for the production of CCL11 in the lung include T lymphocytes, macrophages, eosinophils, smooth muscle, fibroblasts, bronchial epithelial cells, and, most recently, airway parasympathetic neurons (33, 55, 56, 57, 58); our identification of CCL5, CCL11, CCL24, CCL26 mRNA and CCL27 in primary HPAECs and demonstration of their capacity to release these chemokines following serum starvation are hence novel. The absence of any eosinophil survival effect of conditioned medium from alternative endothelial sources, including coronary artery, aorta, and umbilical vein also suggests a degree of heterogeneity and possible specificity in the generation of these chemokines. However, we have not studied the capacity of endothelial cells from alternative sites to express mRNA for CC chemokines or their capacity to secrete these agents following stimulation, for example with IL-13, IL-1β, or TNF-α. In a previous detailed description of cytokine and chemokine expression in HPAECs, HCAECs, and HUVECs (18), these cell types were all shown to express transcripts for IL-1α, IL-5, IL-8, MCP-1, and GM-CSF. In addition, there was weak expression of IL-6 and CCL5 in HCAECs and HPAECs, respectively, following TNF-α activation. Furthermore, the chemokine CCL27 has previously only been shown to be expressed by keratinocytes (59). Of interest, IL-8 levels were also found to be increased in the HPAEC-CM compared with HUVEC-CM. Although IL-8 itself does not influence eosinophil survival (32), it is a potent chemotactic, priming, and activation stimulus in neutrophils (60), suggesting that increased IL-8 release by pulmonary artery endothelial cells may recruit additional inflammatory cells and therefore amplify inflammation. This is supported by reports showing that when neutrophils and eosinophils are coincubated with IL-8, there is enhanced eosinophil migration across Matrigel-coated Transwell inserts (22). It is also interesting to speculate whether recognized microvascular abnormalities detected in asthma and other pulmonary disorders (61) may further modulate the local production of eosinophil-active chemokines from endothelial cells.

These observations suggest that HPAECs have a previously unrecognized capacity to elaborate and secrete CCR3 chemokines, in particular CCL11, and for this agent to operate as a survival agent in human eosinophils. This extends the previously recognized repertoire of actions for CCL11 (eotaxin-1) and suggests a mechanism for the aberrant survival of eosinophils in airway allergic inflammation.

We thank Drs. P. Jose, D. Sexton, and A. Alcami for helpful discussions.

The authors have no financial conflict of interest.

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 Asthma-U.K., Biotechnology and Biological Sciences Research Council, Aventis, GlaxoSmithKline, and the Wellcome Trust.

3

Abbreviations used in this paper: HPAEC, human pulmonary artery endothelial cell; HAEC, human aortic endothelial cell; HCAEC, human coronary artery endothelial cell; CM, conditioned medium; CTACK, cutaneous T cell-attracting chemokine; PI, propidium iodide; CrmE, cytokine response modifier E.

1
Bousquet, J., P. Chane, J. Y. Lacoste, G. Barneon, N. Ghavanian, I. Enander, P. Venge, S. Ahlstedt, J. Simony-Lafontaine, P. Godard.
1990
. Eosinophilic inflammation in asthma.
N. Engl. J. Med.
323
:
1033
-1039.
2
Woolley, K. L., P. G. Gibson, K. Carty, A. J. Wilson, S. H. Twaddell, M. J. Woolley.
1996
. Eosinophil apoptosis and the resolution of airway inflammation in asthma.
Am. J. Respir. Crit. Care Med.
54
:
237
-243.
3
Kankaanranta, H., M. A. Lindsay, M. A. Giembycz, X. Zhang, E. Moilanen, P. J. Barnes.
2000
. Delayed eosinophil apoptosis in asthma.
J. Allergy Clin. Immunol.
106
:
77
-83.
4
Gleich, G. J..
2000
. Mechanisms of eosinophil-associated inflammation.
J. Allergy Clin. Immunol.
105
:
651
-663.
5
Collins, P. D., S. Marleau, D. A. Griffiths-Johnson, P. J. Jose, T. J. Williams.
1995
. Cooperation between interleukin-5 and the chemokine eotaxin to induce eosinophil accumulation in vivo.
J. Exp. Med.
182
:
1169
-1174.
6
Clutterbuck, E. J., E. M. Hirst, C. J. Sanderson.
1989
. Human interleukin-5 (IL-5) regulates the production of eosinophils in human bone marrow cultures: comparison and interaction with IL-1, IL-3, IL-6 and GM-CSF.
Blood
73
:
1504
-1512.
7
Palframan, R. T., P. D. Collins, T. J. Williams, S. M. Rankin.
1998
. Eotaxin induces a rapid release of eosinophils and their progenitors from the bone marrow.
Blood
91
:
2240
-2248.
8
Buckley, C. D., E. A. Ross, H. M. McGettrick, C. E. Osbourne, O. Haworth, C. Schmutz, P. C. Stone, M. Salmon, N. M. Matharu, R. K. Vohra, et al
2006
. Identification of a phenotypically and functionally distinct population of long-lived neutrophils in a model of reverse endothelial migration.
J. Leukoc. Biol.
79
:
303
-311.
9
Erjefalt, J. S., C. G. A. Persson.
2000
. New aspects of degranulation and fates of airway mucosal eosinophils.
Am. J. Respir. Crit. Care Med.
161
:
2074
-2085.
10
Tsuyuki, S., C. Bertrand, F. Erard, A. Trifilieff, J. Tsuyuki, M. Wesp, G. P. Anderson, A. J. Coyle.
1995
. Activation of the Fas receptor on lung eosinophils leads to apoptosis and the resolution of eosinophilic inflammation of the airways.
J. Clin. Invest.
96
:
2924
-2931.
11
Walsh, G. M., D. W. Sexton, M. G. Blaylock, C. M. Convery.
1999
. Resting and cytokine-stimulated human small airway epithelial cells recognize and engulf apoptotic eosinophils.
Blood
94
:
2827
-2835.
12
Stern, M., L. Meagher, J. Savill, C. Haslett.
1992
. Apoptosis in human eosinophils: programmed cell death in the eosinophil leads to phagocytosis by macrophages and is modulated by IL-5.
J. Immunol.
148
:
3543
-3549.
13
Simon, H.-U., S. Yousefi, C. Schranz, A. Schapowal, C. Bachert, K. Blaser.
1997
. Direct demonstration of delayed eosinophil apoptosis as a mechanism causing tissue eosinophilia.
J. Immunol.
158
:
3902
-3908.
14
Ying, S., Q. Meng, L. Taborda-Barata, A. B. Kay.
1997
. Association of apoptosis of neutrophils and eosinophils and their ingestion by macrophages with resolution of the allergen-induced cutaneous late phase response in atopic human subjects.
Proc. Assoc. Am. Physicians
109
:
42
-50.
15
Hu, M., X. Lin, Q. Du, E. J. Miller, P. Wang, H. H. Simms.
2005
. Regulation of polymorphonuclear leukocyte apoptosis: role of lung endothelium-epithelium bilayer transmigration.
Am. J. Physiol.
288
:
L266
-L274.
16
Yamamoto, H., J. B. Sedgwick, R. F. Vrtis, W. W. Busse.
2000
. The effect of transendothelial migration on eosinophil function.
Am. J. Respir. Cell Mol. Biol.
23
:
379
-388.
17
Bochner, B. S., F. W. Luscinskas, M. A. Gimbrone, Jr, W. Newman, S. A. Sterbinsky, C. P. Derse-Anthony, D. Klunk, R. P. Schleimer.
1991
. Adhesion of human basophils, eosinophils, and neutrophils to interleukin 1-activated human vascular endothelial cells: contributions of endothelial cell adhesion molecules.
J. Exp. Med.
173
:
1553
-1557.
18
Krishnaswamy, G., J. Kelley, L. Yerra, J. K. Smith, D. S. Chi.
1999
. Human endothelium as a source of multifunctional cytokines: molecular regulation and possible role in human disease.
J. Interferon Cytokine Res.
19
:
91
-104.
19
Haslett, C., L. A. Guthrie, M. M. Kopaniak, R. B. Johnston, Jr, P. M. Henson.
1985
. Modulation of multiple neutrophil functions by preparative methods or trace concentrations of bacterial lipopolysaccharide.
Am. J. Pathol.
119
:
101
-110.
20
Cowburn, A. S., K. Cadwallader, B. Reed, N. Farahi, E. R. Chilvers.
2002
. Role of PI3-kinase dependent BAD phosphorylation and altered transcription in cytokine-mediated neutrophil survival.
Blood
100
:
2607
-2616.
21
Morrell, N. W., P. D. Upton, S. Kotecha, A. Huntley, M. H. Yacoub, J. M. Polak, J. Wharton.
1999
. Angiotensin II activates MAPK and stimulates growth of human pulmonary artery smooth muscle via AT1 receptors.
Am. J. Physiol.
277
:
L440
-L448.
22
Kikuchi, I., S. Kikuchi, T. Kobayashi, K. Hagiwara, Y. Sakamoto, M. Kanazawa, M. Nagata.
2006
. Eosinophil trans-basement membrane migration induced by interleukin-8 and neutrophils.
Am. J. Respir. Cell Mol. Biol.
34
:
760
-765.
23
Qu, J., A. M. Condliffe, M. Lawson, R. J. Plevin, R. A. Riemersma, G. R. Barclay, D. B. McClelland, E. R. Chilvers.
1995
. Lack of effect of recombinant platelet-derived growth factor on human neutrophil function.
J. Immunol.
154
:
4133
-4141.
24
Kitchen, E., A. G. Rossi, A. M. Condliffe, C. Haslett, E. R. Chilvers.
1996
. Demonstration of reversible priming of human neutrophils using platelet-activating factor.
Blood
88
:
4330
-4337.
25
Yamaguchi, Y., T. Suda, S. Ohta, K. Tominaga, Y. Miura, T. Kasahara.
1991
. Analysis of the survival of mature human eosinophils: interleukin-5 prevents apoptosis in mature human eosinophils.
Blood
78
:
2542
-2547.
26
Krishnaswamy, G., J. K. Smith, R. Mukkamala, K. Hall, W. Joyner, L. Yerra, D. S. Chi.
1998
. Multifunctional cytokine expression by human coronary endothelium and regulation by monokines and glucocorticoids.
Microvasc. Res.
55
:
189
-200.
27
Levi-Schaffer, F., P. Lacy, N. J. Severs, T. M. Newman, J. North, B. Gomperts, A. B. Kay, R. Moqbel.
1995
. Association of granulocyte-macrophage colony-stimulating factor with the crystalloid granules of human eosinophils.
Blood
85
:
2579
-2586.
28
Tai, P. C., L. Sun, C. J. Spry.
1991
. Effects of IL-5, granulocyte/macrophage colony-stimulating factor (GM-CSF) and IL-3 on the survival of human blood eosinophils in vitro.
Clin. Exp. Immunol.
85
:
312
-316.
29
Matsumoto, R., M. Hirashima, H. Kita, G. J. Gleich.
2002
. Biological activities of ecalectin: a novel eosinophil-activating factor.
J. Immunol.
168
:
1961
-1967.
30
Okada, S., J. B. Hagan, M. Kato, J. L. Bankers-Fulbright, L. W. Hunt, G. J. Gleich, H. Kita.
1998
. Lidocaine and its analogues inhibit IL-5-mediated survival and activation of human eosinophils.
J. Immunol.
160
:
4010
-4017.
31
Parry, C. M., J. P. Simas, V. P. Smith, C. A. Stewart, A. C. Minson, S. Efstathiou, A. Alcami.
2000
. A broad spectrum secreted chemokine binding protein encoded by a herpesvirus.
J. Exp. Med.
191
:
573
-578.
32
Petering, H., O. Gotze, D. Kimmig, R. Smolarski, A. Kapp, J. Elsner.
1999
. The biological role of interleukin-8: functional analysis and expression of CXCR1 and CXCR2 on human eosinophils.
Blood
93
:
694
-702.
33
Fryer, A. D., L. H. Stein, Z. Nie, D. E. Curtis, C. M. Evans, S. T. Hodgson, P. J. Jose, K. E. Belmonte, E. Fitch, D. B. Jacoby.
2006
. Neuronal eotaxin and the effects of CCR3 antagonist on airway hyperreactivity and M2 receptor dysfunction.
J. Clin. Invest.
116
:
228
-236.
34
Saraiva, M., A. Alcami.
2001
. CrmE, a novel soluble tumor necrosis factor receptor encoded by poxviruses.
J. Virol.
75
:
226
-233.
35
Iversen, P. O., D. Robinson, S. Ying, Q. Meng, A. B. Kay, I. Clark-Lewis, A. F. Lopez.
1997
. The GM-CSF analogue E21R induces apoptosis of normal and activated eosinophils.
Am. J. Respir. Crit. Care Med.
156
:
1628
-1632.
36
Uller, L., M. Anderson, L. Grieff, C. G. A. Persson, J. S. Erjefalt.
2004
. Occurrence of apoptosis, secondary necrosis, and cytolysis in eosinophilic nasal polyps.
Am. J. Respir. Crit. Care Med.
170
:
742
-747.
37
Erjefalt, J. S., L. Uller, M. Malm-Erjefalt, C. G. Persson.
2004
. Rapid and efficient clearance of airway tissue granulocytes through transepithelial migration.
Thorax
59
:
136
-143.
38
Her, E., J. Frazer, K. F. Austin, W. F. Owen, Jr.
1991
. Eosinophil haematopoietins antagonize the programmed cell death of eosinophils: cytokine and glucocorticoid effects on eosinophils maintained by endothelial cell-conditioned medium.
J. Clin. Invest.
88
:
1982
-1987.
39
Luttmann, W., B. Knoechel, M. Foerster, H. Matthys, J. C. Virchow, Jr, C. Kroegel.
1996
. Activation of human eosinophils by IL-13: induction of CD69 surface antigen, its relationship to messenger RNA expression, and promotion of cellular viability.
J. Immunol.
157
:
1678
-1683.
40
Hoontrakoon, R., H. W. Chu, S. J. Gardai, S. E. Wenzel, P. McDonald, V. A. Fadok, P. M. Henson, D. L. Bratton.
2002
. Interleukin-15 inhibits spontaneous apoptosis in human eosinophils via autocrine production of granulocyte-colony stimulating factor and nuclear factor-κB activation.
Am. J. Respir. Crit. Care Med.
26
:
404
-412.
41
Lee, E., T. Robertson, J. Smith, S. Kilfeather.
2000
. Leukotriene receptor antagonists and synthesis inhibitors reverse survival in eosinophils of asthmatic individuals.
Am. J. Respir. Crit. Care Med.
161
:
1881
-1886.
42
Alam, R., K. Pazdrak, S. Stafford, P. Forsythe.
1995
. The interleukin-5/receptor interaction activates Lyn and Jak2 tyrosine kinases and propagates signals via the Ras-Raf-1-MAP kinase and the Jak-STAT pathways in eosinophils.
Int. Arch. Allergy Immunol.
107
:
226
-227.
43
Fujihara, S., C. Ward, I. Dransfield, R. T. Hay, I. J. Uings, B. Hayes, S. N. Farrow, C. Haslett, A. G. Rossi.
2002
. Inhibition of nuclear factor-κB activation un-masks the ability of TNF-α to induce human eosinophil apoptosis.
Eur. J. Immunol.
32
:
457
-466.
44
Pazdrak, K., B. Olszewska-Pazdrak, S. Stafford, R. P. Garofalo, R. Alam.
1998
. Lyn, Jak2, and Raf-1 kinases are critical for the antiapoptotic effect of interleukin 5, whereas only Raf-1 kinase is essential for eosinophil activation and degranulation.
J. Exp. Med.
188
:
421
-429.
45
Farahi, N., A. S. Cowburn, A. G. Rossi, E. R. Chilvers.
2004
. Eating their way out of trouble: selective uptake of apoptotic eosinophils by bronchial epithelial cells.
Clin. Exp. Allergy
34
:
1503
-1506.
46
Leckie, M. J., A. ten Brinke, J. Khan, Z. Diamant, B. J. O’Connor, C. M. Walls, A. K. Mathur, H. C. Cowley, K. F. Chung, R. Djukanovic, et al
2000
. Effects of an interleukin-5 blocking monoclonal antibody on eosinophils, airway hyper-responsiveness, and the late asthmatic response.
Lancet
356
:
2144
-2148.
47
Bryan, S. A., B. J. O’Connor, S. Matti, M. J. Leckie, V. Kanabar, J. Khan, S. J. Warrington, L. Renzetti, A. Rames, J. A. Bock, et al
2000
. Effects of recombinant human interleukin-12 on eosinophils, airway hyper-responsiveness, and the late asthmatic response.
Lancet
356
:
2114
-2116.
48
Wedi, B., U. Raap, U. Lewrick, A. Kapp.
1998
. IL-4-induced apoptosis in peripheral blood eosinophils.
J. Allergy Clin. Immunol.
102
:
1013
-1020.
49
Shinagawa, K., A. Trifilieff, G. P. Anderson.
2003
. Involvement of CCR3-reactive chemokines in eosinophil survival.
Int. Arch. Allergy Immunol.
130
:
150
-157.
50
Conroy, D. M., T. J. Williams.
2001
. Eotaxin and the attraction of eosinophils to the asthmatic lung.
Respir. Res.
2
:
150
-156.
51
Jose, P. J., D. A. Griffiths-Johnson, P. D. Collins, D. T. Walsh, R. Moqbel, N. F. Totty, O. Truong, J. J. Hsuan, T. J. Williams.
1994
. Eotaxin: a potent chemoattractant cytokine detected in a guinea pig model of allergic airways inflammation.
J. Exp. Med.
179
:
881
-887.
52
Menzies-Gow, A., S. Ying, I. Sabroe, V. L. Stubbs, D. Soler, T. J. Williams, A. B. Kay.
2002
. Eotaxin (CCL11) and eotaxin-2 (CCL24) induce recruitment of eosinophils, basophils, neutrophils and macrophages as well as features of early- and late-phase allergic reactions following cutaneous injection in human atopic and nonatopic volunteers.
J. Immunol.
169
:
2712
-2718.
53
Joubert, P., S. Lajoie-Kadoch, I. Labonte, A. S. Gounni, K. Maghni, V. Wellemans, J. Chakir, M. Laviolette, Q. Hamid, B. Lamkhioued.
2005
. CCR3 expression and function in asthmatic airway smooth muscle cells.
J. Immunol.
175
:
2702
-2708.
54
Dent, G., C. Hadjicharalambous, T. Yoshikawa, R. L. Handy, J. Powell, I. K. Anderson, R. Louis, D. E. Davies, R. Djukanovic.
2004
. Contribution of eotaxin-1 to eosinophil chemotactic activity of moderate and severe asthmatic sputum.
Am. J. Respir. Crit. Care Med.
169
:
1110
-1117.
55
Ghaffar, O., Q. Hamid, P. M. Renzi, Z. Allakhverdi, S. Molet, J. C. Hogg, S. A. Shore, A. D. Luster, B. Lamkhioued.
1999
. Constitutive and cytokine-stimulated expression of eotaxin by human airway smooth muscle cells.
Am. J. Respir. Crit. Care Med.
159
:
1933
-1942.
56
Hoeck, J., M. Woisetschlager.
2001
. STAT6 mediates eotaxin-1 expression in IL-4 or TNF-α-induced fibroblasts.
J. Immunol.
166
:
4507
-4515.
57
Ying, S., D. S. Robinson, Q. Meng, J. Rottman, R. Kennedy, D. J. Ringler, C. R. Mackay, B. L. Daugherty, M. S. Springer, S. R. Durham, et al
1997
. Enhanced expression of eotaxin and CCR3 mRNA and protein in atopic asthma: association with airway hyperresponsiveness and predominant co-localization of eotaxin mRNA to bronchial epithelial and endothelial cells.
Eur. J. Immunol.
27
:
3507
-3516.
58
Chan, V. J., J. K. Burgess, J. C. Ratoff, B. J. O’Connor, A. Greenough, I. H. Lee, S. J. Hirst.
2006
. Extracellular matrix regulates enhanced cotaxin expression in asthmatic airway smooth muscle cells.
Am. J. Respir. Crit. Care Med.
174
:
379
-385.
59
Morales, J., B. Homey, A. P. Vicari, S. Hudak, E. Oldham, J. Hedrick, R. Orozco, N. G. Copeland, N. A. Jenkins, L. M. McEvoy, A. Zlotnik.
1999
. CTACK, a skin-associated chemokine that preferentially attracts skin-homing memory T cells.
Proc. Natl. Acad. Sci. USA
96
:
14470
-14475.
60
Teran, L. M., M. P. Carroll, A. J. Frew, A. E. Redington, D. E. Davies, I. Lindley, P. H. Howarth, M. K. Church, S. T. Holgate.
1996
. Leukocyte recruitment after local endobronchial allergen challenge in asthma: relationship to procedure and to airway interleukin-8 release.
Am. J. Respir. Crit. Care Med.
154
:
469
-476.
61
McDonald, D. M..
2001
. Angiogenesis and remodelling of airway vasculature in chronic inflammation.
Am. J. Respir. Crit. Care Med.
164
:
S39
-S45.