Proc. Natl. Acad. Sci. USA Vol. 77, No. 12, pp. 7440-7443, December 1980 Medical Sciences

Identification of human eosinophil lysophospholipase as the constituent of Charcot-Leyden crystals ( B) PETER F. WELLER*t, EDWARD J. GOETZL*t, AND K. FRANK AUSTEN* Department of Medicine, Harvard Medical School, and Department of Rheumatology and Immunology, Brigham and Women's Hospital, Boston, Massachusetts 02115; and tHoward Hughes Medical Institute Laboratories for the Study of Immunological Diseases ofHarvard Medical School, Boston, Massachusetts 02115 Contributed by K. Frank Austen, September 15,1980

ABSTRACT Since the initial descriptions of Charcot-Ley- protein in the large granule of the eosinophil (12). The intra- den crystals more than 100 years ago, the presence of these cellular source of the eosinophil CLC protein remains uncertain, slender, di yramidal c stals in human tissues and biologic although nuclear (9), granular (10), and cytoplasmic (12) origins fluids has become a hallmark of eosinophilic leukocye infil- tration, especially in association with allergic and helminthic have been proposed. Further, no biochemical activity or bio- diseases. The formation of these crystals in vitro after disruption logic function has been established for the protein composing of human eosinophils, but not of other cell types, in hypotonic the crystals (11). The findings that chromatographically puri- saline or detergent established the eosinophil as the unique fied, homogeneous eosinophil lysophospholipase (lysolecithin cellular source of the crystalline protein. Charcot- yden acylhydrolase, EC 3.1.1.5) and CLC protein express a common crystals have now been found to express lysophospholipase ac- tivity (lysolecithin acylhydrolase, EC 3.1.1.5), and the solubilized enzymatic activity, exhibit the same physicochemical charac- Charcot-Leyden crystal protein presents a single stained protein teristics of size and charge, and form crystals of identical band that is coincident with the lysophospholipase activity morphology establish human eosinophil lysophospholipase as eluted from replicate gels on alkaline polyacrylamide gel elec- the constituent of CLC. trophoresis. On sodium dodecyl sul ate/polyacrylamide gel electrophoresis, the solubilized Charcot-Leyden crystal protein migrates with a molecular weight of 17,400, which is compa- MATERIALS AND METHODS rable to that of eosinophil lysophospholipase purified chroma- Materials. Reagents and equipment were obtained from the tographically to homogeneity; further, on combination, the two proteins comigrate as a single staining band. Finally, the chro- sources indicated: L-1-[1-'4C]Palmitoyl matographically purified eosinophil lysophospholipase in hy- (specific activity about 50 mCi/mmol; 1 Ci = 3.7 X 1010 bec- potonic buffer forms dipyramidal crystals morphologically querels) and Aquasol scintillation fluid from New England identical to Charcot-Leyden crystals. The findings that chro- Nuclear; Sephadex G-100, heparin-Sepharose, gel filtration and matographically purified, homogeneous eosinophil lysophos- electrophoresis molecular weight calibration proteins, 6% pholipase and Charcot-Leyden crystal protein express the same enzymatic activity, are of the same size and charge, and form dextran 70 in normal saline (Macrodex), and diatriazoate-Ficoll crystals of identical morphology indicate that human eosinophil (Ficoll-Paque) from Pharmacia; synthetic palmitoyl lyso- lysophospholipase is the constituent of Charcot-Leyden crys- phosphatidylcholine, egg-yolk phosphatidylcholine, iodo- tals. acetamide, and glycine from Sigma; organomercurial agarose (Affi-Gel 501), NaDodSO4, acrylamide, N,N'-methylene- Prominent blood or tissue accumulations of eosinophilic leu- bisacrylamide, N,N,N',N'-tetramethylethylenediamine, urea, kocytes occur in many allergic reactions and helminthic in- dithiothreitol, ammonium persulfate, Coomassie brilliant blue fections as well as in the course of some neoplastic, inflamma- stain, and Bio-Rad protein assay from Bio-Rad; Na2EDTA, tory, and immunodeficiency diseases (1). Charcot-Leyden chloroform, methanol, heptane, isopropanol, Spectrapor dialysis crystals (CLC), which are characteristically long, slender, di- membranes, and Gelman SA ITLC thin-layer chromatography pyramidal crystals, were initially described in the mid-nine- sheets from Fisher; calcium- and magnesium-free Hanks' teenth century (2, 3) and are hallmarks of eosinophil involve- balanced salt solution from Microbiological Associates (Walk- ment in some tissue reactions. CLC may be prominent in spu- ersville, MD); 10,000 Mr retention collodion bags from tum of patients with asthma (3), pulmonary ascariasis (4), and Schleicher & Schuell; Branson model 350 sonifer from Branson tropical eosinophilia (5), in the feces of patients with amebic, Sonic Power; and UM-05 Diaflo membranes from Amicon. Trichuris, or ulcerative colitis (6), and in tissues containing Blood from normal donors provided neutrophils, mononu- eosinophilic granulomas of bone (7) and granulomas associated clear cells, platelets, and erythrocytes (13), and donors with with tissue-invading helminths (8). blood eosinophilias due to trichinosis, filariasis, drug allergies, Disruption of eosinophils either by detergents (9) or by ho- and a hypereosinophilic syndrome provided granulocytes en- mogenization and suspension in hypotonic saline (5) allows for riched in eosinophils (14). Eosinophils from a donor with hyp- the formation of CLC in vitro. Similar disruption of other ereosinophilia (leukocyte count of 28,000-36,000/mm3 of blood classes of leukocytes or of eosinophils derived from nonprimate with 87-94% eosinophils) were used without further purifica- mammalian species fails to result in CLC formation (10). Thus, tion as a source of protein for formation of CLC and for isolation eosinophils, specifically those from primate species, have been of . established as the unique cellular source of this crystalline Assay of Lysophospholipase. The cell preparations were material (11). Human CLC are composed of a single protein, disrupted by sonication and dialyzed against 0.1 M NaCl which differs from the major basic protein, the predominant overnight before assay of cellular lysophospholipase. A 100 to 200-,1 portion of cell extract or chromatographic fraction was The publication costs of this article were defrayed in part by page of 1 con- charge payment. This article must therefore be hereby marked "ad- added to a reaction mixture with a final volume ml, vertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact. Abbreviation: CLC, Charcot-Leyden crystals. 7440 Downloaded by guest on September 24, 2021 Medical Sciences: Weller et al. Proc. Nati. Acad. Sci. USA 77 (1980) 7441

taining 250 nmol of synthetic palmitoyl lysophosphatidylcho- pH 8.9,7.5% acrylamide gels (17) with the modifying addition line and 0.96 nmol of [14C]palmitoyl lysophosphatidylcholine, of 2 mM dithiothreitol to the electrode buffer. Gels were elec- both of which had been sonicated for 2 min at4VC at setting trophoresed at 2.5 mA per gel for 3 hr before use. 5 with a microtip (Branson model 350 sonifier) in 0.1 M Tris- HCI, pH 7.5/2 mM EDTA. Five hundred micrograms of RESULTS phosphatidylcholine, sonicated in buffer for 9min as above, was For the same numbers of disrupted human peripheral blood included in the assay of chromatographic fractions. The reac- leukocytes, lysophospholipase activity in the eosinophil-en- tion mixture was incubated at 370C for 1 hr and was then ex- riched granulocytes was 8-fold greater than that in the neu- tracted with acidified isopropanol/heptane (15) to partition the trophil-enriched granulocytes and 3-fold greater than that of released fatty acid into the upper organic layer, from which a the mononuclear leukocytes (Fig. 1). Relatively trivial activity sample was taken and its radioactivity was measured in Aquasol was present in platelets and erythrocytes. (14). The identification and quantitation of the released fatty In a preliminary experiment, 200 Ml of pelleted CLC con- acid was confirmed by extracting replicate incubation mixtures tained 184 units of lysophospholipase activity when assayed with chloroform/methanol/water (10:10:9, vol/vol) and re- directly. To determine if this enzyme activity was a major solving the extracted lipids by thin-layer chromatography on component of the crystal, lyophilized CLC were solubilized in silicic acid with a neutral lipid solvent system in comparison 0.1 M Tris-HCI, pH 7.5/2 mM EDTA, and 40-jg quantities of with purified standards (14). One unit of enzyme activity protein, as assessed by the Bio-Rad assay using human IgG as represents 1 nmol of fatty acid liberated per hour at 37"C, at a reference standard, were subjected to alkaline disc gel elec- pH 7.5. trophoresis in two replicate gels. In the gel stained with Coo- Purification of Eosinophil Lysophospholipase. Approxi- massie brilliant blue, a single stained protein band was present. mately 1 X 109 leukocytes (94% eosinophils, 6% neutrophils) The parallel gel was sliced into 2-mm sections and each section were disrupted by sonication in 0.05 M Tris-HCI, pH 8.5/2 mM was macerated in 500,ul of 0.1 M Tris-HCl, pH 7.5/2 mM EDTA. After the cells were centrifuged at 750 X g for 10 min, EDTA/2 mM dithiothreitol and dialyzed overnight against the the supernatant fluid was subjected to gel filtration on Sephadex same buffer at 40C with 6000-8000 Mr cut-off dialysis mem- G-100 in 0.05 M Tris.HCI, pH 8.5/2 mM EDTA. The fractions branes. Of the lysophospholipase activity recovered from eluted containing the peak of enzyme activity, which eluted at about slices of this gel, 88% was in a 6-mm region coincident with the 66% of bed volume, were pooled and applied to a column of stained protein band (Fig. 2). organomercurial agarose equilibrated in 0.1 M Tris-HCI, pH Eosinophil lysophospholipase, purified to homogeneity as 7.5/2 mM EDTA. Lysophospholipase eluted early with a gra- judged by presentation of a single stained protein band on al- dient from 0 to 0.35 M dithiothreitol in 0.1 M Tris-HCI, pH kaline disc gel electrophoresis and NaDodSO4/polyacrylamide 7.5/2 mM EDTA/0. 15 M NaCl. The peak lysophospholipase- gel electrophoresis of 40 ,ug each, was subjected to Na- containing fractions were adjusted to a conductivity of 4 mS and DodSO4/polyacrylamide gel electrophoresis in parallel and in applied to a heparin-Sepharose column equilibrated with 60 concert with CLC protein. Each protein migrated with the mM Tris-HCI, pH 7.5/1.2 mM EDTA. Lysophospholipase ac- same RF and in combination they comigrated to give a single tivity, which appeared in the fall-through fractions, was ho- band (Fig. 3). The molecular weight of CLC (Fig. 4), using 10 mogeneous, demonstrating a single Coomassie brilliant blue ,g of CLC for each of six determinations, was 17,400 : 300 staining band on both NaDodSO4/10% polyacrylamide gel (SEM), which is the same as the molecular weight of human electrophoresis and 7.5% polyacrylamide gel electrophoresis eosinophil lysophospholipase, 17,200 + 200 (SEM) (n = 9). at pH 8.9. A solution of 900,g of purified eosinophil lysophospholipase Preparation of CLC. After dextran-induced sedimentation in 500 Al of 0.1 M Tris-HCl, pH 7.5/2 mM EDTA/2 mM di- of erythrocytes, 1.64 X 109 leukocytes from a patient with hy- thiothreitol, was made hypotonic by addition of 1 ml of 0.15% pereosinophilic syndrome were washed twice in calcium and NaCl and concentrated at 40C to original volume by nega- magnesium-free Hanks' balanced salt solution and residual tive-pressure ultrafiltration on a 10,000 Mr collodion bag. erythrocytes were lysed by the addition of 0.15% (wt/vol) NaCl. Crystals with a morphology identical to that of CLC formed The eosinophil-rich leukocytes were sedimented at 700 X g for within minutes of concentration (Fig. 5). 2 min at room temperature and disrupted by sonication on ice for 2 min at an output of 7 with a microtip in 45 ml of 0.15 % NaCl. Particulate material was removed by sequential cen- Eosinophils trifugation at 2000 X g for 15 min at 4°C and then at 32,000 5 X 106 cells = X g for 30 min at 4°C. The supernatant was held overnight at (73 !t 6% eos) 4-fold by positive-pressure ultra- Neutrophils 4°C and then concentrated 5 X 106 cells nl5 filtration with a UM-05 Diaflo membrane to facilitate crystal (3 4 1% eos) formation (5). Crystals were collected and washed once in cold Mononuclear cells n = distilled H20 by centrifugation at 2000 X g for 20 min at 4°C. 5 X 106 cells After microscopic confirmation of the characteristic crystalline Platelets n= 1 morphology, the sample was lyophilized. 2 X 108 cells Analytical Methods. NaDodSO4/polyacrylamide gel elec- Erythrocytes n trophoresis was performed in 10% polyacrylamide gels (16). 109 cells I 8 M urea were treated with 10 mM dithiothreitol Samples in 0 20 40 60 80 100 at 56°C for 5 min, followed by alkylation in 10 mM iodoacet- Lysophospholipase activity, units amide at 56°C for 15 min. NaDodSO4 was added to a concen- FIG. 1. Lysophospholipase activity of human blood cells dis- of 1% (wt/vol), and the samples were incubated at tration rupted by sonication and dialyzed against 0.1 M NaCl before assay. 1000C for 10 min. Protein standards included phosphorylase Enzyme activity is presented as the mean (+SEM) units for prepa- b, bovine serum albumin, ovalbumin, carbonic anhydrase, rations from n different donors; 1 unit is equivalent to 1 nmol of fatty soybean trypsin inhibitor, and a-lactalbumin. Analytical acid liberated per hr at 370C from palmitoyl lysophosphatidylcho- polyacrylamide disc gel electrophoresis was performed with line. Downloaded by guest on September 24, 2021 7442 Medical Sciences: Weller et al. Proc. Natl. Acad. Sci. USA 77 (1980)

l lII77l 1 1 1 I lI 105

4.) 40[- OA bo _1 CA 'Z-

_c _

_ .S X_ ) I c- .v. TI CLC

c- -r LA

11-_ 0.2 0.4 0.6 0.8 1.0 RF FIG. 4. NaDodSO4/polyacrylamide gel electrophoresis in 10%o gels of reduced and alkylated samples of CLC protein and standard N fNj I proteins, including I I LtII IIIIIIIJJI ILLL I I ILI..ILI a-lactalbumin (LA), soybean trypsin inhibitor 21 1 1 1 i 21'1 22 (TI), carbonic anhydrase (CA), and ovalbumin (OA). Relative mi- SAlice( gration values (RF) are presented as a ratio ofthe distance migrated FIG. 2. Alkaline disc gel electrophoresis of solubilized CLC for each protein relative to the migration distance of bromphenol blue protein derived from human eosinophils. Protein was applied to tracking dye. parallel gels in 40-ag portions; one gel was stained with Coomassie brilliant blue and the other was sliced and eluted for assay of lyso- based on quantitating the release of radiolabeled fatty acid from phospholipase activity. The anode was at the right and the dye front [14C]palmitoyl lysophosphatidylcholine (14). Human eosino- was marked by an ink stab. phils express 8-fold more lysophospholipase activity than do neutrophils and 3fold more than mononuclear leukocytes (Fig. DISCUSSION 1). Thus, in the human, and possibly in rats (19) and mice (20), The observation of prominent lysophospholipase activity as- the eosinophil is preferentially endowed with the enzymatic sociated with the eosinophil was made with rat leukocytes on capability to deacylate lysophospholipids. the basis of the histochemical precipitation of (18) and the ti- During the course of purifying lysophospholipase from the tration of (19) fatty acid liberated from the lysolec- human eosinophil (14), the similarity of physicochemical ithin. Human eosinophils manifest more lysophospholipase properties between this enzyme and the protein composing activity than other leukocytes in a sensitive and specific assay CLC suggested that the two might be related. CLC prepared from human eosinophils by ultrasonic disruption of the cells in hypotonic saline were found to express prominent lysophos- pholipase activity. Alkaline disc gel electrophoresis of CLC protein yielded a single band of Coomassie brilliant blue staining protein that was coincident with lysophospholipase activity recovered from a parallel gel (Fig. 2), indicating that the activity was inherent to the CLC protein and not merely

A B

-ab- _ I I _w0_ I I 11 ,

FIG. 3. NaDodSO4polyacrylamide gel electrophoresis in 10% gels of 45 gg of purified eosinophil lysophospholipase (left), the combination of 15 Mg of purified eosinophil lysophospholipase and 20 Mg of eosinophil-derived CLC protein (center), and 40 ,ug of eo- sinophil-derived CLC protein (right). All samples were reduced and FIG. 5. CLC prepared from human eosinophils disrupted by alkylated. The bromphenol blue dye front was marked with a stab of sonication in 0.15% NaCl (A) and from purified eosinophil lyso- ink. phospholipase (B). (X1400.) Downloaded by guest on September 24, 2021 Medical Sciences: Weller et al. Proc. Natl. Acad. Sci. USA 77 (1980) 7443

adsorbed to the crystal. Moreover, on electrophoresis under 1. Weller, P. F. & Goetzl, E. J. (1979) Adv. Immunol. 27, 339- dissociating conditions, after reduction and alkylation in the 371. presence of NaDodSO4 detergent, CLC protein migrated as a 2. Charcot, J. M. & Robin, C. (1853) Mem. Soc. Biol. 5, 44-50. single band with the same mobility as purified eosinophil ly- 3. Leyden, E. (1872) Virchow's Arch. Pathol. Anat. 54,324-344. 4. Beaver, P. C. & Danaraj, T. J. (1958) Am. J. Trop. Med. 7, sophospholipase; indeed, the two proteins comigrated on a 100-111. single gel (Fig. 3). Additional confirmation that eosinophil ly- 5. Archer, F. T. & Blackwood, A. (1965) J. Exp. Med. 122, 173- sophospholipase was the sole protein constituent of CLC was 180. obtained by the demonstration that purified enzyme in hypo- 6. Beaver, P. C. (1960) in Proceedings of the Sixth International tonic buffer formed dipyramidal crystals morphologically Congress on Tropical Medicine and Maladies, Lisbon, Portugal, identical to CLC (Fig. 5). Vol. 3, pp. 419-434. Chemical knowledge of CLC was limited until Gleich and 7. Ayres, W. W. & Silliphant, W. M. (1958) Am. J. Clin. Pathol. 30, his associates found CLC to consist of a 13,000 Mr acidic protein 323-327. whose molecular weight and amino acid composition distin- 8. Beaver, P. C. (1962) Bull. Soc. Pathol. Exot. 55,471-475. guished it from the 9300 Mr major basic protein, the principal 9. Ayres, W. W. & Starkey, N. M. (1950) Blood 5,254-266. constituent of eosinophil large granules (12, 21). In the present 10. El-Hashimi, W. (1971) Am. J. Pathol. 65,311-324. protein 11. Beeson, P. B. & Bass, D. A. (1977) The Eosinophil (Saunders, New study, molecular weights of the CLC and the purified York), pp. 39-42. lysophospholipase are similar, being 17,400 and 17,200, re- 12. Gleich, G. J., Loehering, D. A., Mann, V. G. & Maldonado, J. E. spectively. The greater molecular weight of CLC in the present (1976) J. Clin. Invest. 57,633-640. study as compared to the molecular weights in the literature 13. Boyum, A. (1968) Scand. J. Clin. Lab. Invest. Suppl. 97, 21, (12, 21) is not due to aberrant migration of eosinophil lyso- 77-89. phospholipase, because analysis of the electrophoretic mobility 14. Weller, P. F., Wasserman, S. I. & Austen, K. F. (1980) in The on NaDodSO4/polyacrylamide gel electrophoresis performed Eosinophil in Health and Disease, eds. Mahmoud, A. A. F. & at four different concentrations of acrylamide, as suggested by Austen, K. F. (Grune & Stratton, New York), pp. 115-128. Ferguson (22), gave the same result. The introduction of ad- 15. Dole, V. P. & Meinertz, H. (1960) J. Biol. Chem. 235, 2595- 2599. ditional molecular weight markers, the a and chains of he- 16. Weber, K. & Osborn, M. (1969) J. Biol. Chem. 244, 4406- moglobin (about 15,000 Mr) (16), confirmed the larger size of 4412. CLC and lysophospholipase. Some (5, 23), but not all (24), 17. Maurer, H. R. (1971) Disc Electrophoresis and Related Tech- studies of CLC have found increased quantities of zinc associ- niques of Polyacrylamide Gel Electrophoresis (de Gruyter, New ated with CLC, and the sulfhydryl dependence of lysophos- York), p. 44. pholipase (18) allows for the possible binding of zinc ions by 18. Ottolenghi, A., Pickett, J. P. & Greene, W. B. (1967) J. Histochem. sulfhydryl groups of the cysteine residues (12, 23). However, Cytochem. 14,907-914. the purification of lysophospholipase in the presence of 2 mM 19. Ottolenghi, A. (1969) Lipids 5, 531-538. EDTA and the ability of purified enzyme to form CLC in the 20. Ottolenghi, A. (1973) Lipids 9, 426-428. presence of EDTA suggest that zinc is not an integral compo- 21. Ackerman, S. J., Loegering, D. & Gleich, G. J. (1980) Fed. Proc. nent of the crystals. Fed. Am. Soc. Exp. Biol. 39,805 (abstr.). in 22. Ferguson, K. A. (1964) Metabolism 13, 985-1002. CLC persist sputum, feces, and tissues of patients with 23. Buddecke, E., Essellier, A. F. & Marti, H. R. (1956) Z. Physiol. eosinophil inflammatory processes and are resistant to proteo- Chem. 305, 203-206. lytic digestion by chymotrypsin, trypsin, and pepsin (25). The 24. Hornung, M. (1962) Proc. Soc. Exp. Biol. Med. 110, 119-124. retention of active lysophospholipase activity by CLC derived 25. Dawe, C. I. & Williams, W. L. (1953) Anat. Rec. 116, 53- from human eosinophils raises the possibility that, even after 73. the disintegration of eosinophils in inflammatory foci, the re- 26. Hirata, F., Corcoran, B. A., Venkatasubramanian, K., Schiffmann, sulting CLC that are formed may represent an enzymatically E. & Axelrod, J. (1979) Proc. Natl. Acad. Sci. USA 76, 2640- active legacy that continues to function to degrade lysophos- 2643. pholipids. Because lysophospholipids are generated in in- 27. Keller, R. (1962) Helv. Physiol. Pharmacol. Acta 20, 66. 28. Reman, F. C., Demel, R. A., DeGier, J., van Deenen, L. L. M., flammatory foci (26, 27) and may have a range of concentra- Eibl, H. & Westphal, 0. (1969) Chem. Phys. Lipids 3, 221- tion-dependent cytotoxic and noncytolytic effects on diverse 233. cells (28-32), CLC formation and persistence in human tissues 29. Joist, J. H., Dolezel, G., Cucuianu, M. P., Nishizawa, E. E. & and fluids may represent more than a curious crystalline artifact Mustard, J. F. (1977) Blood 49, 101-112. and may have biological function. 30. Martin, T. W. & Lagunoff, D. (1979) Nature (London) 279, 250-252. 31. Anderson, W. B. & Jaworski, C. J. (1977) Arch. Biochem. Biophys. The authors gratefully acknowledge the technical assistance of Ms. 180,374-383. Mary Gulash and Mr. David Bach. This work was supported by Grants 32. Shier, W. T., Baldwin, J. H., Nilsen-Hamilton, M., Hamilton, R. AI-07722, HL-19777, and RR-05669 from the National Institutes of T. & Thanassi, N. M. (1976) Proc. Natl. Acad. Sci. USA 73, Health. 1586-1590. Downloaded by guest on September 24, 2021