Eosinophil Interaction with Antibody- Coated, Non-Phagocytosable Surfaces: Changes in Cell Surface Proteins

J. Cell Sci. 42, 367-378 (1980) 367 Printed in Great Britain © Company of Biologists Limited igSo

EOSINOPHIL INTERACTION WITH ANTIBODY- COATED, NON-PHAGOCYTOSABLE SURFACES: CHANGES IN CELL SURFACE PROTEINS

KAREEN J. I. THORNE, RHONDA C. OLIVER AND AUDREY M. GLAUERT Strangeways Research Laboratory, Worts Causeway, Cambridge, England, CBi 4.RN

SUMMARY Plasma membrane changes during the interaction of human eosinophils with large, antibody- coated, non-phagocytosable surfaces have been investigated in a model system. Human peripheral blood eosinophils were incubated with layers of agar into which tetanus toxoid and human anti-tetanus immunoglobulin, together with an eosinophil chemotactic factor (ECF), were incorporated. Changes in organization of the eosinophil plasma membrane proteins during interaction with the agar layer were detected by lactoperoxidase-catalysed iodination with [125]iodide. A protein of apparent mol. wt 55000 became newly accessible on the eosinophil surface as a specific consequence of interaction with antigen-antibody complexes in the agar layer. This protein appeared in the early attachment phase of the interaction which preceded extracellular degranulation. Cytochalasin D enhanced its appearance, while Mg2+-deficiency prevented it. A second newly accessible protein of apparent mol. wt 58000 was blocked when ECF was present and may therefore be a receptor for ECF. Other proteins of apparent mol. wt 68000 and 46000 newly appeared at the surface of eosinophils even after incubation in suspension, apparently as a consequence of the rapid cycling of membrane components which occurs in eosinophils.

INTRODUCTION The discovery that eosinophils appear to play a major role in host resistance to schistosomula of Schistosoma mansoni (Butterworth et al. 1975; Butterworth, 1977) has drawn attention to the importance of the eosinophil as a cytotoxic cell. When human peripheral blood eosinophils encounter antibody-coated schistosomula they attach to and flatten down closely onto the schistosomulum surface in an apparent attempt to phagocytose it (Glauert & Butterworth, 1977; Glauert, Butterworth, Sturrock & Houba, 1978). This frustrated phagocytosis is followed by release of the eosinophil granule contents onto the surface of the schistosomulum and by subsequent damage both to the surface tegument and to the underlying muscle layers (Glauert et al. 1978). In order to investigate the mechanism of eosinophil attachment to and degranulation onto large non-phagocytosable surfaces we have developed a model system in which eosinophils interact with human antigen-antibody complexes in agar layers (Glauert, Oliver & Thorne, 1980). The antigen-antibody complex used is tetanus toxoid with human anti-tetanus antiserum. In addition the eosinophil chemotactic 368 K. J. I. Thome, R. C. Oliver and A. M. Glauert factor, ala-gly-ser-glu, (ECF) (Kay, Stechschulte & Austen, 1971; Goetzl & Austen, 1975) is incorporated into the agar layer. Eosinophils flatten down closely onto this layer and adhere more intimately to it than do neutrophils (Glauert et al. 1980). Both cell types then degranulate, as detected by the extracellular release of collagenase, peroxidase and /?-glucuronidase (Glauert et al. 1980). The behaviour of eosinophils on the agar layer therefore resembles their behaviour on the surface of the schistosomulum. This model is now being used to elucidate the biochemical aspects of the interaction of eosinophils, in comparison with neutrophils, with non-phagocytosable surfaces. In the present study we have looked for changes in organization of the plasma membrane proteins of the eosinophil during the interaction. The plasma membrane proteins accessible on the cell surface have been detected by lactoperoxidase-catalysed iodination with [125I]iodide. Increase or decrease in surface accessibility has been related to individual stages in the interaction of the eosinophil with the non-phagocytosable surface.

MATERIALS AND METHODS Eosinophil and neutroph.il preparation Eosinophils and neutrophils were prepared from normal human peripheral blood by a modification of the method of Vadas et al. (1979). The details have been described elsewhere (Thome, Glauert, Svvennsen & Franks, 1979) but a brief outline of the method is given here. Heparinized venous blood (20 ml) was depleted of erythrocytes by sedimentation with 0-35 % (w/v) methyl cellulose in phosphate-buffered saline. Granulocytes were prepared from the leucocyte-rich supernatant by centrifugation on a Ficoll/Hypaque gradient. These were separated into neutrophils and eosinophils on a 12-ml Metrizamide gradient, 20-25 % (w/v) in Hanks buffered salts solution (HBSS) containing o-i % (w/v) gelatin and 15 /tg/ml DNAse I (Sigma Chemical Co., Kingston upon Thames, KT2 7BH). After centrifugation for 45 min at 400 g the upper white band of neutrophils and the lower red band of eosinophils were collected and the Metrizamide removed by washing with HBSS containing o-i % (w/v) gelatin.

Attachment to agar layers Thin layers of agar, containing IO~5M ala-gly-ser-glu tetrapeptide, an eosinophil chemotactic factor (ECF) (Uniscience Ltd., Cambridge CB5 8BA), and tetanus toxoid (Lister Institute, Elstree, Herts., WD6 3AY) were prepared on glass coverslips as described by Glauert et al. (1980). Immediately before use the agar layers were treated with 10 times diluted human anti- tetanus immunoglobulin (Lister Institute) for 45 min at 37 °C. Excess immunoglobulin was then removed and the layers washed thoroughly with HBSS. Purified eosinophils (70-100%) or neutrophils (90-100%) in HBSS, containing o-i % (w/v) bovine serum albumin (BSA) or o-i % (w/v) gelatin were then allowed to settle onto and attach to the agar layers, supported on glass coverslips in Falcon 3008 multi-cell tissue culture plates. After incubation for 15- 60 min at 37 °C, the HBSS containing the unattached cells was removed and the layers were washed once with HBSS, containing BSA or gelatin. The number of cells adhering to the layers was estimated by counting the number of unattached cells present in the removed HBSS and the wash. After 15 min about 20 % of the added eosinophils were attached. The unattached cells were then removed from the washes by centrifugation and the residual supernatant was assayed for released granule enzymes.

[125I]Iodination of cell surface proteins Eosinophils and neutrophils were pretreated with non-radioactive potassium iodide to block pre-existing accessible tyrosine residues in the resting cells. The cells were incubated in Eosinophil surface proteins 369 suspension in 1 ml HBSS with 20 /tg/ml lactoperoxidase (Sigma Chemical Co.), 004 units of Type V glucose oxidase (Sigma Chemical Co.), and o-i mil potassium iodide for 10 min at room temperature. They were then washed 3 times with HBSS containing o-i % gelatin. The cells were attached to agar layers on glass coverslips as described above. After removal of the unattached cells the cell monolayer was treated with 1 ml HBSS with 20 /tg/ml lactoperoxidase, 0-04 units of glucose oxidase and about 100 /tCi L125I]sodium iodide for 10 min at room temperature. The iodinating solution was then removed and the cells on agar were washed 3 times with HBSS containing o-i % gelatin. The 125I-labelled cells were solubilized with 0-2 ml of digestion mixture (4 % (w/v) sodium dodecyl sulphate (SDS), 20 % (v/v) mercaptoethanol and 0-09% (w/v) bromphenol blue in aqueous solution) at 100 °C for 1 min. The digest was then separated by polyacrylamide gel electrophoresis on rods of 75 % polyacrylamide in o-i % SDS and o-i M Tris/bicine, pH 83. After electrophoresis the gels were sliced into i-mm-long segments and counted in a Packard PGD auto gamma counter. The gels were calibrated with the following standard proteins: phosphorylase a (mol. wt iooooo), transferrin (mol. wt 78000), bovine serum albumin (mol. wt 68000), ovalbumin (mol. wt 42000), carbonic anhydrase (mol. wt 29000), soya bean trypsin inhibitor (mol. wt 21000) and cytochrome c (mol. wt 13000).

Electron microscopy Eosinophils and neutrophils were incubated in 0-2 ml of HBSS and gelatin on agar layers on Araldite instead of on glass coverslips. The cells were then fixed in situ by the addition of 20 n\ of 25 % glutaraldehyde and incubation was continued at 37 °C for 30 min. The glutaraldehyde was then removed and the cells washed with o-i M cacodylate buffer pH 7-2 containing 3 mM CaCL. The fixed cells on the agar layers were then processed for electron microscopy as described earlier (Glauert et al. 1980).

Enzyme assays Peroxidase (E.C. 1.11.1.7) was assayed by the method of Migler & DeChatelet (1978) using p-phenylenediamine as hydrogen donor. /?-glucuronidase (E.C. 3.2.1.31) was assayed employing phenolphthalein /?-D-glucuronide as substrate. The enzyme was incubated at 37 °C for 2 h with 1 mM phenolphthalein /?-D-glucuronide and o-i M sodium acetate buffer, pH 4-5. The reaction was stopped by the addition of 2 vol. of 1 M sodium carbonate and the amount of free phenolphthalein determined from its optical density at 555 nm.

RESULTS Attachment of eosinophils and neutrophils to agar layers Eosinophils and neutrophils differed markedly in the nature of their attachment to agar layers containing the chemotactic factor (ECF) and antigen-antibody complexes (Figs. 1, 2). After 30 min eosinophils had flattened down closely onto the surface of the agar layer while neutrophils made many contacts of limited area with intervening regions where the cell was not in contact with the agar. As already described (Glauert et al. 1980) attachment was usually complete in 30 min and this was followed by degranulation. Release of granule enzymes from neutrophils commenced as soon as the cells were added to the agar layers, while release of eosinophil enzymes began to be detectable only after a lag of about 1 h.

Changes in plasma membrane proteins during interaction with agar layers The surface proteins of human eosinophils which were accessible to lactoperoxidase- catalysed iodination proved to be more numerous than those previously observed on 370 K. J. I. Thome, R. C. Oliver and A. M. Glauert Eosinophil surface proteins

10 20 30 40 5 3 B 2 1 - - f - 1 7 o 3 4 X 10 20 30 40 50 £

jf 2

10 20 30 40 50 3

10 20 30 40 50 60

I 100 68 42 29 21 13 Mol. wtX 10'3 Fig. 3. Lactoperoxidase-catalysed iodination of newly accessible cell-surface proteins. After iodination the proteins were separated by SDS-PAGE in 7*5 % polyacrylamide. The gels were sliced into i-mm segments and the radioactivity of each slice plotted against the distance along the gel. A, B, C, cells on agar layers containing antigen- antibody complex and ECF. A, eosinophils incubated for 15 min; B, eosinophils incubated for 60 min; c, neutrophils incubated for 15 min; D, erythrocytes incubated in suspension for 15 min. rabbit peritoneal exudate neutrophils (Thome, Oliver & Lackie, 1977 a). In order, therefore, to detect changes in the organization of membrane proteins pre-existing accessible tyrosine residues on the cell surface were first iodinated with unlabelled sodium iodide; the cells were then allowed to interact with the agar layers, unattached cells were removed, and the attached cells were iodinated with [125I]iodide for detection

Figs. 1, 2. Electron micrographs of eosinophils and neutrophils adhering to agar layers containing io"° M ECF and antigen-antibody complexes. Bar lines, 1 /im. Fig. 1. After 60 min incubation an eosinophil has made close parallel contact with the agar surface (ag). Fig. 2. After 30 min incubation a neutrophil is only loosely attached to the agar surface (ag) by broad pseudopodia. 372 K. J. I. Thome, R. C. Oliver and A. M. Glauert of new groups on the cell surface. The cells retained their biological activity as judged by their ability to release peroxidase during a subsequent 2-h incubation with the agar layers (30% released from untreated cells; 36% from iodinated cells). They also retained their ability to phagocytose Trypanosoma dionisii (Thorne et al. 1979). After iodination the radioactive proteins were separated by polyacrylamide gel electrophoresis in SDS (Fig. 3). The results are expressed as radioactivity/io° cells attached

Fig. 4. Lactoperoxidase-catalysed iodination of newly accessible cell surface proteins. (Experimental procedure as Fig. 3.) Effect of ECF and antigen-antibody complexes. Eosinophils incubated for 15 min (A) in suspension for 15 min with ECF and antigen- antibody complex, (B) on agar layers containing antigen-antibody complex, and (c) on agar layers containing ECF. to the agar layers. All experiments were performed at least twice, and the experiment shown in Fig. 3A was performed 6 times, with consistent results. After eosinophils had interacted with antigen-antibody complexes on agar layers for 15 min, 3 new proteins appeared at the cell surface (Fig. 3 A). These proteins were designated 1 (apparent mol. wt 68000), 3 (apparent mol. wt 55000) and 4 (apparent mol. wt 46000). The considerable amount of radioactive material moving Eosinophil surface proteins 373 at the front is believed to be lipid. The appearance of the new proteins at the surface of the eosinophil appeared to represent an early response to the antibody-coated agar layer, since if the cells were left for 60 min on agar very little new protein was detected (Fig. 3 u). Neutrophil labelling was much less than the labelling of eosinophils (Fig. 3 c). To eliminate the possibility that the new proteins detected on the eosinophil surface were from contaminating erythrocytes the labelling pattern of erythrocytes was determined (Fig. 3 D). It was found to be quite different from that of eosinophils. Nor could the labelling be attributed to the bovine serum albumin in the medium since the same results were obtained when albumin was replaced with gelatin. In different experiments the radioactivity of the main peak 1 varied from 3 to 9 x io4 cpm/106 cells attached. Eosinophils incubated in suspension for 15 min with ECF and antigen-antibody complexes produced only proteins 1 and 4 (Fig. 4A). These proteins of mol. wt 68000 and 46000 appeared under all conditions tested, in the presence and absence of antigen-antibody complexes and in the presence and absence of ECF. If eosinophils were incubated with agar layers in the absence of ECF or in the absence of antigen-antibody complexes some cells attached but the pattern of labelling of these cells was altered. In the absence of ECF an additional protein, 2 (apparent mol. wt 58000) could be detected (Fig. 413). This was not seen when ECF was present. In the absence of antigen-antibody complexes very little protein 3 was found (Fig. 4c).

Effect of inhibitors Inhibitors of adherence and of degranulation were tested for their effect on the modification of the eosinophil surface proteins. The influence of these agents on the interaction with antibody-coated agar layers, containing ECF, is summarized in Table 1. Little effect on the total number of cells attached was observed, although the nature of the attachment could be affected. For example, some cells may adhere much more intimately than others (see Glauert et al. 1980). Since extracellular release of the granule enzyme peroxidase is believed to be consequent upon attachment to the agar layer, the amount of peroxidase found in the medium after 2 h incubation is given both as a percentage of the total cell content and as a percentage of that in the attached cells. Cytochalasins inhibit both close adherence of cells to glass or plastic surfaces (Weiss, 1972) and phagocytosis (see review by Allison, 1973), but enhance extracellular degranulation in the presence of aggregated immunoglobulin (Zurier, Hoffstein & Weissmann, 1973). Cytochalasin D was used in the present experiments in preference to cytochalasin B since it affects microfilaments without also inhibiting hexose transport (Miranda, Godman, Deitch & Tanenbaum, 1974; Miranda, Godman & Tanenbaum, 1974). Attachment in the presence of cytochalasin D (Fig. 5 c) enhanced the appearance of protein 3 and also revealed protein 2, which had previously been detected only in the absence of ECF (Fig. 4). Cytochalasin D increased the amount of enzyme released on agar layers nearly 3-fold. cAMP inhibits degranulation in neutrophils (Ignarro, Paddock & George, 1974), mast cells (Kaliner & Austen, 1974) and platelets (Maclntyre et al. 1977). In our 374 K. J. I. Thome, R. C. Oliver and A. M. Glauert

Table i. Effect of inhibitors on attachment of eosinophils to antibody-coated layers and on release of peroxidase

Peroxidase release

Attachment Total cells Attached cells

1 1 > 1 + Agent, Control, + Agent, Control, + Agent, Control, Agent 0//o % %±S.D. %±S.D. % 0//o Ratio

Cytochalasin D, 55 69 35 ±1° l6 + 2 64 23 28 io/tg/ml 005 mM dibutyryl 35 42 I-2±0'3 5'2 ±2-4 3 13 025 cAMP + o-os niM theophylline 2+ Ca -free + o-1 mM 65 5i 7-5 + 2-9 5-1 ±2-9 11 10 I-I EGTA 2+ Mg -free+ 1 m^i 43 5° 41 ±13 49 ±05 9'5 10 10 EDTA 2+ 2+ Mg - and Ca - 24 28 0 2-6±o-s 0 9 0 free + 1 mM EDTA

Eosinophils were incubated for 2 hwith agar layers containing antigen-antibody and ECF, in the presence and absence of the listed agents.

agar model system dibutyryl cAMP and theophylline reduced peroxidase release from adherent eosinophils by three quarters (Table 1), but had no effect on the appearance of new proteins on the eosinophil surface (Fig. 5 B), suggesting that their appearance precedes and is not a consequence of degranulation. Ca2+-depletion with EGTA in a Mg2+-containing medium not only had little effect on eosinophil attachment and degranulation (Table 1) but also had no detectable effect on the appearance of new surface proteins (Fig. 6B). Mg2+-depletion alone also had little effect, but elimination of both cations inhibited peroxidase release completely (Table 1). However, it was the absence of Mg2+ which prevented the appearance of protein 3 (Fig. 6c).

DISCUSSION The technique used in the present paper for observing changes in cell surface proteins is lactoperoxidase-catalysed iodination with [125I]iodide and hydrogen peroxide. Changes in platelet membrane proteins occurring during lipopolysaccharide- induced aggregation have been detected by a double-labelling method (Thorne, Oliver, Maclntyre & Gordon, 19776) where resting platelets were labelled with [125I]iodide and stimulated platelets were labelled with [131I]iodide and the ratio of isotope in each membrane protein determined. This method worked well for rabbit platelets since relatively few proteins on the control platelet surface were accessible to iodination. Human peripheral blood eosinophils and neutrophils, however, had Eosinophil surface proteins 375 too many proteins on the cell surface for clear separation and quantitation. For this reason the cells were pretreated instead with unlabelled iodide to block the exposed tyrosine residues. Human eosinophils in suspension in medium appeared to be in a dynamic state and to be constantly bringing more of 2 proteins of apparent mol. wt 68000 (protein 1)

10 20 30 40 50 12 8 O

X E

10 20 30 40 50 mm Fig. 5. Lactoperoxidase-catalysed iodination of newly accessible cell surface proteins. (Experimental procedure as Fig. 3.) Effect of cAMP and cytochalasin D. Eosinophils incubated for 15 min on agar layers containing ECF and antigen-antibody complexes. A, no inhibitors; B, 0-05 mM theophylline; C, io/tg/ml cytochalasin D. and 46000 (protein 4) to the surface. In this they differed from rabbit peritoneal neutrophils, in which a new protein of mol. wt 150000 was detectable at the cell surface only when the rate of membrane cycling was accelerated by phagocytosis followed by exocytosis (Thorne et al. 1977a). Eosinophils brought new proteins to the surface even when they were not presented with a phagocytosable particle. Sanderson & Thomas (1978) showed by time-lapse photomicrography that rat peritoneal eosinophils show rapid membrane movement. The newly accessible 376 K. J. I. Thome, R. C. Oliver and A. M. Glauert protein on the human eosinophil surface of mol. wt 68000 (protein 1) is unlikely to be serum albumin from the medium since it was also detected when the experiments were performed in gelatin. An additional newly accessible protein of mol. wt 55000 (protein 3) appeared to be a specific consequence of interaction of the eosinophil with an antibody-coated, non- phagocytosable surface and may therefore be associated with, or be part of, the Fc

Fig. 6. Lactoperoxidase-catalysed iodination of newly accessible cell surface proteins. (Experimental procedure as Fig. 3.) Eosinophils incubated for 15 min with agar layers containing ECF and antigen-antibody complexes. A, complete system; B, Caa+-free medium containing o-i mM EGTA; C, Mga+-free medium containing 1 mM EDTA. receptor in the eosinophil membrane. The position of protein 3 in the membrane may be partly controlled by links with intracellular microfilaments, since its appearance at the cell surface is enhanced by cytochalasin D. The protein appears to be involved in an early antibody- and Mg2+-dependent event preceding degranulation. It is of interest that incubation in the presence of cytochalasin modifies the membrane changes induced in eosinophils during interaction with agar layers, while Eosinophil surface proteins 377 cAMP has no effect, since both agents inhibit the killing of schistosomula by eosinophils (David et al. 1977). Presumably cytochalasin interferes with the killing of schistosomula by modifying the attachment of the eosinophil, while cAMP inhibits by preventing the subsequent release of the toxic basic protein from the eosinophil granules (Butterworth et al. 1979). If the chemotactic factor (ECF) was omitted from the agar model system a second new protein of apparent mol. wt 58000 (protein 2) was exposed during interaction with the antibody-coated agar layers. It is tempting to postulate that this might be a receptor for ECF and that in the complete system it is covered by ECF. Boswell, Austen & Goetzl (1976) have evidence that the eosinophil-reeeptor for the tetrapeptide contains a hydrophobic domain which recognizes the N-terminal valine or alanine, a hydrogen-bonding domain which recognizes serine, and an ionic domain which recognizes the C-terminal glutamate. The newly accessible proteins on the cell surface of the eosinophil may be divided into 2 groups. The first group of proteins, 1 and 4, seems to be constantly reappearing proteins on the eosinophil surface and to appear even on the cell in suspension. These proteins may be brought to the surface by the rapid cycling of membrane components resulting from endocytosis followed by exocytosis, of which the eosinophil is capable. The second group of proteins, 2 and 3, appears on the eosinophil surface as a consequence of its interaction with a non-phagocytosable antigen-antibody complex, and seems therefore to result from rearrangement of pre-existing plasma membrane proteins. Protein 2 appears to be associated with the receptor for the chemotactic factor (ECF), while the appearance of the protein 3 is an early event in the attachment of the eosinophil to non-phagocytosable antigen-antibody complexes, and appears to be involved in the initial antibody-dependent interaction which precedes extracellular secretion of lysosomal granule enzymes. We thank Mr R. A. Parker for technical assistance, and The Wellcome Trust for financial support and for providing the Philips 201C electron microscope.

REFERENCES ALLISON, A. C. (1973). The role of microfilaments and microtubules in cell movement, endocytosis and exocytosis. Ciba Fdn Symp. 14, 109-148. BOSWELL, R. N., AUSTEN, K. F. & GOETZL, E. J. (1976). A chemotactic receptor for val (ala)-gly-ser-glu on human eosinophil polymorphonuclear leukocytes. Immunol. Commun. 5, 469-479. BUTTERWORTH, A. E. (1977). The eosinophil and its role in immunity to helminth infection. Curr. Top. Microbiol. Immun. 77, 127-168. BUTTERWORTH, A. E., STURROCK, R. F., HOUBA, V., MAHMOUD, A. A. F., SHER, A. & REES, P. H. (1975). Eosinophils as mediators of antibody-dependent damage to schistosomula. Nature, Lond. 256, 727-729. BUTTERWORTH, A. E., WASSOM, D. L., GLEICH, G. J., LOEGERING, D. A. & DAVID, J. R. (1979). Damage to schistosomula of Schistosoma mansoni induced directly by eosinophil major basic protein.,?. Immun. 122, 221-229. DAVID, J. R., BUTTERWORTH, A. E., REMOLD, H. G., DAVID, P. H., HOUBA, V. & STURROCK, R. F. (1977). Antibody-dependent eosinophil-mediated damage to 51Cr-labelled schistosomula of Schistosoma mansoni: effect of metabolic inhibitors, and other agents which alter cell function. J. Immun. 118, 2221-2229.

25 CEL42 378 K. J. I. Thome, R. C. Oliver and A. M. Glauert GLAUERT, A. M. & BUTTERWORTH, A. E. (1977). Morphological evidence for the ability of eosinophils to damage antibody-coated schistosomula. Trans. R. Soc. trop. Med. Hyg. 71, 392-395- GLAUERT, A. M., BUTTERWORTH, A. E., STURROCK, R. F. & HOUBA, V. (1978). The mechanism of antibody-dependent, eosinophil-mediated damage to schistosomula of Schistosoma mansoni in vitro: a study by phase-contrast and electron microscopy. J. Cell Sci. 34,173-192. GLAUERT, A. M., OLIVER, R. C. & THORNE, K. J. I. (1980). The interaction of human eosinophils and neutrophils with non-phagocytosable surfaces: a model for studying cell-mediated immunity in schistosomiasis. Parasitology (In Press). GOETZL, E. J. & AUSTEN, K. F. (1975). Purification and synthesis of eosinophilotactic tetra- peptides of human lung tissue: identification as eosinophil chemotactic factor of anaphylaxis. Proc. natn. Acad. Sci. U.S.A. 72, 4123-4127. IGNARRO, L. J., PADDOCK, R. J. & GEORGE, W. J. (1974). Hormonal control of neutrophil lysosomal enzyme release: effect of epinephrine on adenosine 3',s'-monophosphate. Science, N.Y. 183, 855-857. KAY, A. B., STECHSCHULTE, D. J. & AUSTEN, K. F. (1971). An eosinophil leukocyte chemotactic factor of anaphylaxis. J. exp. Med. 133, 602-619. KALINER, M. & AUSTEN, K. F. (1974). Cyclic AMP, ATP and reversed anaphylactic histamine release from rat mast cells. J. Ivimun. 112, 664-674. MACINTYRE, D. E., ALLEN, A. P., THORNE, K. J. I., GLAUERT, A. M. & GORDON, J. L. (1977). Endotoxin-induced platelet aggregation and secretion. I. Morphological changes and pharmacological effects. J. Cell Sci. 28, 211-223. MIGLER, R. & DECHATELET, L. R. (1978). Human eosinophilic peroxidase; biochemical characterization. Biochem. Med. 19, 16-26. MIRANDA, A. F., GODMAN, G. C, DEITCH, A. D. & TANENBAUM, S. W. (1974). Action of cytochalasin D on cells of established lines. I. Early events. J. Cell Biol. 6i, 481-500. MIRANDA, A. F.( GODMAN, G. C. & TANENBAUM, S. W. (1974). Action of cytochalasin D on cells of established lines. II. Cortex and microfilaments. J. Cell Biol. 62, 406-423. SANDERSON, C. J. & THOMAS, J. A. (1978). A comparison of the cytotoxic activity of eosinophils and other cells by "chromium release and time lapse microcinematography. Immunology 34, 771-780. THORNE, K. J. I., GLAUERT, A. M., SWENNSEN, R. J. & FRANKS, D. (1979). Phagocytosis and killing of Trypanosoma dionisii by human neutrophils, eosinophils and monocytes. Para- sitology 79, 367-379- THORNE, K. J. I., OLIVER, R. C. & LACKIE, J. (1977a). Changes in the surface properties of rabbit polymorphonuclear leucocytes, induced by bacteria and bacterial endotoxin. J. Cell Sci. 27, 213-225. THORNE, K.J.I., OLIVER, R. C, MACINTYRE, D. E. & GORDON, J. L. (1977 b). Endotoxin- induced platelet aggregation and secretion. II. Changes in plasma membrane proteins, jf. Cell Sci. 28, 225-236. VADAS, M. A., DAVID, J. R., BUTTERWORTH, A. E., PISANI, N. T. & SIONGOK, T. A. (1979). A new method for the purification of human eosinophils and neutrophils, and a comparison of the ability of these cells to damage schistosomula of Schistosoma mansoni. J. Immun. 122, 1228-1236. WEISS, L. (1972). Studies on cellular adhesion in tisssue culture. XII. Some effects of cytochalasins and colchicine. Expl Cell Res. 74, 21-26. ZURIER, R. B., HOFFSTEIN, S. & WEISSMANN, G. (1973). Cytochalasin B: effect on lysosomal enzyme release from human leukocytes. Proc. natn. Acad. Sci. U.S.A. 70, 844-848. (Received 2 August 1979)