Eosinophil granules function extracellularly as receptor-mediated secretory organelles

Abstract

Intracellular granules in several types of leukocytes contain preformed proteins whose secretions contribute to immune and inflammatory functions of leukocytes, including eosinophils, cells notably associated with asthma, allergic inflammation, and helminthic infections. Cytokines and chemokines typically elicit extracellular secretion of granule proteins by engaging receptors expressed externally on the plasma membranes of cells, including eosinophils. Eosinophil granules, in addition to being intracellular organelles, are found as intact membrane-bound structures extracellularly in tissue sites of eosinophil-associated diseases. Neither the secretory capacities of cell-free eosinophil granules nor the presence of functional cytokine and chemokine receptors on membranes of leukocyte granules have been recognized. Here, we show that granules of human eosinophils express membrane receptors for a cytokine, IFN-gamma, and G protein-coupled membrane receptors for a chemokine, eotaxin, and that these receptors function by activating signal-transducing pathways within granules to elicit secretion from within granules. Capacities of intracellular granule organelles to function autonomously outside of eosinophils as independent, ligand-responsive, secretion-competent structures constitute a novel postcytolytic mechanism for regulated secretion of eosinophil granule proteins that may contribute to eosinophil-mediated inflammation and immunomodulation.

Fig. 1.

Extracellular eosinophil granules are recognized in tissues in association with diverse disorders. (A) Extracellular eosinophil granules (arrows) are present in a skin lesion of a subject with hypereosinophilic syndrome. (Scale bar: 1 μm.) (B) Cell-free eosinophil granules (arrowhead and higher magnified in Bi) and intact eosinophils (arrows and higher magnified in Bii) were identified in a skin biopsy of a patient with verrucous cancer. Preparations were stained with H&E and examined by light microscopy. [Scale bar: 17 μm, 3.6 μm (Bi), and 6 μm (Bii).]

Fig. 2.

Extracellular eosinophil granules function as cell-free secretion-competent organelles in response to IFN-γ and eotaxin. (A) Eosinophil granules secrete ECP in response to IFN-γ (solid line and squares) and eotaxin (dashed line and solid triangles). Secreted ECP levels, means of duplicates ± SD, are ECP levels from stimulated granules minus ECP levels from unstimulated granules, as assayed after 1 h by ELISA. Data from one experiment are representative of three, each of which demonstrated significant dose-dependent secretion elicited by eotaxin and IFN-γ (by ANOVA). *P < 0.05 vs. IFN 1-ng/ml stimulated samples. (B) AO loaded granules (arrows) respond 0, 10, 13, 32, 35, and 92 sec after IFN-γ stimulation with intense transient fluorescent flashes indicative of release of monomeric AO. Fluorescence intensity is pseudocolored, with red representing the greatest intensity as indicated by the scale color. (Scale bar: 3 μm.)

Fig. 3.

Extracellular eosinophil granules express IFN-γ and CCR3 receptors on their membranes. By flow cytometry of isolated non–membrane-permeabilized granules, granule surface membranes expressed extracellular domains of the IFN-γR α chain (A) and the CCR3 receptor for eotaxin (B) but not the nonligand (carboxy)-terminal intracellular domain of CCR3 (C). Shaded histograms and solid lines represent staining with control Ab and mAb or pAb-specific Abs, respectively. Data are from one experiment, representative of three.

Fig. 4.

IFN-γ– and eotaxin-elicited ECP secretion by cell-free granules is mediated by intragranular signaling pathways. Pretreatment of eosinophil isolated granules with indicated concentrations of genistein, a tyrosine kinase inhibitor (A), SB203580 and SB 202190, both p38 MAPK inhibitors (B), and calphostin C, a PKC inhibitor (C), with each inhibited ECP secretion elicited by IFN-γ. (D) PI3K inhibitor, LY294002, did not inhibit ECP secretion elicited by IFN-γ. Eotaxin-induced ECP secretion was inhibited by pertussis toxin (PTX) (E), SB203580 and SB 202190 (F), calphostin C (G), and LY294002 (H). Eosinophil granules were pretreated with inhibitors for 15 min and stimulated with IFN-γ (200 ng/ml) or eotaxin (100 ng/ml) for 1 h. Secreted ECP levels, means of duplicates ± SD, are ECP levels from stimulated granules minus ECP levels from unstimulated granules as assayed by ELISA. Data, one experiment representative of three, were analyzed by one-way ANOVA, followed by the Newman-Keuls test. + and * represent P < 0.05 compared with nonstimulated and cytokine/chemokine-stimulated granules, respectively. (I) Isolated granules contain PKC and p-38 MAPK. (J) Eotaxin (100 ng/ml, 1 min) induced phosphorylation of p-38 MAPK on granules. t, total; p, phosphorylated; NS, not stimulated; Eot, eotaxin-stimulated. (K) IFN-γ (200 ng/ml, 5 min) induced tyrosine phosphorylation of a granule protein, and genistein (100 μM) inhibited phosphorylation. Loading controls were blotted for granule MPB. pp tyr, phosphorylated tyrosine residues; NS, not stimulated; IFN, IFN-γ–stimulated; Gen, genistein.

Fig. 5.

BFA inhibits ECP secretion from IFN-γ– (A) and eotaxin- (B) stimulated eosinophil granules. Secreted ECP levels, means of duplicates ± SD, are ECP levels from stimulated granules minus ECP levels from unstimulated granules, as assayed after 1 h by ELISA. Data, one experiment representative of three, were analyzed by one-way ANOVA, followed by the Newman-Keuls test. + and * represent P < 0.05 compared with nonstimulated and cytokine/chemokine-stimulated granules, respectively. (C) BFA promotes the collapse of membranotubular networks within granules forming electron-dense membranolipid deposits (arrows, inset). (Scale bar: 200 nm.) Data are from one experiment, representative of three.

Fig. 6.

Isolated eosinophil granule cytokine secretion upon IFN-γ stimulation. IFN-γ dose dependently induced release of IL-4 and IL-6 but not IL-13 after 1 h. Cytokine levels in unstimulated samples were undetectable. Open, closed, and hatched columns represent stimulation with 100, 200, and 400 ng/ml IFN-γ, respectively. Data represent the means of duplicates ± SD. Results, one experiment representative of three, were analyzed by one-way ANOVA, followed by the Newman-Keuls test. *P < 0.05 compared with nonstimulated granules.