Cytoskeletal changes as an early event in hydrogen - Europe PMC

Br. J. exp. Path. (I986) 67, I05-II2

Cytoskeletal changes as an early event in hydrogen peroxide-induced cell injury: a study in A549 cells *Ganesh Raghu, tLiliane Striker, *John Harlan, Striker

tAllen

Gown and tGary

Departments of *Medicine and tPathology, University of Washington, Seattle, Washington 98195 and tMetabolic Diseases Branch, National Institute of Arthritis, Diabetes, Digestive, and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20205 Received for publication io June I985 Accepted for publication 2 7 August I 9 8 5

Summary. Hydrogen peroxide (H202) and other oxygen metabolites have been implicated in the pathogenesis of cell and tissue injury. The nature of the injury occurring in cells exposed to oxygen metabolites is unknown. A549 cells, derived from human lung carcinoma, were exposed to glucose-glucose oxidase or hydrogen peroxide in vitro. The distribution of actin and cytokeratin filaments, as well as 5chromium (5 'Cr) release and trypan blue dye exclusion were assessed. Both glucose-glucose oxidase and H202 resulted in changes which were timeand dose-dependent. Alterations in the cytoskeleton were detected by immunofluorescence microscopy at two hours, at which time the cells excluded trypan blue dye, while 5 'Cr release and trypan blue uptake first occurred at 8 h and required a five-fold greater concentration of glucose oxidase. The addition of catalase to glucose-glucose oxidase or H202, or inactivation of glucose oxidase by boiling, abrogated the injury. Therefore, one of the early targets of H202induced cell injury may be the cytoskeleton.

Keywords: A549 cells, hydrogen peroxide, cytoskeleton, 5'chromium release

Oxygen metabolites have been demonstrated not by superoxide dismutase, suggesting that to mediate endothelial, epithelial and interH202 was the responsible factor (Sacks et al. stitial cell injury in both intact organs and in I978; Suttorp & Simon I982; Weiss et al. cultured cells (Balin et al. I976; Martin et al. I98I). The susceptibility to lysis of some I 98 I; Parsons et al. I 984; Weiss et al. I 98 I). tumor cells was shown to be linked to the Endothelial cell monolayers labelled with H202 generated by neutrophils or macro51chromium (5 'Cr) have been shown to phages, and was dependent on the activity of release 5'Cr after addition of activated poly- glutathione redox cycle (Nathan et al. I 98 I). morphonuclear leukocytes (Weiss et al. The importance of the glutathione redox I98I; Yamada et al. I98I). Hydrogen per- cycle as an antioxidant mechanism in other oxide (H202) may be inactivated by catalase cultured cells has been established (Harlan et or by the glutathione peroxidase-glutathione al. I984; Meister I983; Roos et al. 1979). reductase system. Cells damaged by oxygen This study was undertaken to examine the metabolites were inhibited by catalase, but early cellular events in H202-mediated Correspondence: Ganesh Raghu, M.D., Assistant Professor of Medicine, Division of Respiratory Diseases and Critical Care, Department of Medicine, RM-I2, University of Washington, Seattle, WA 98I95, USA. I05

G. Raghu et al. Io6 injury to A549 cells. The cytoskeleton was removed in 50 4u1 at intervals to determine examined because we had previously found specific 5'Cr release as follows: that changes in cell shape preceded "Cr ct/min test - ct/min control release, cell detachment, or trypan blue ct/min maximum - ct/min control x 100 uptake following endothelial cell injury (Harlan et al. I982). In this study, changes in Ct/min, counts per minute; control, spontacytoskeleton were found to occur earlier neous 5'Cr release in medium alone; maxithan either 51Cr release or trypan blue mum, 5 Cr release with I% Triton X-ioo uptake. Since the cytoskeleton contributes to (New England Nuclear, Boston, MA). the maintenance of normal cell shape, juncSpontaneous 5 Cr release in control tional complexes, and adherence to cell monolayers was I I% and 25% of maximum substrate, changes in them may be impor- after 8 and 24 hours incubation respectively. tant in the pathogenesis of cell damage Maximum 51Cr release was determined by (Lazarides I980; I982; Schliwa & van Bler- incubation in I% Triton X-ioo. kom I 98 I; Weber & Osborne I 982). Immunofluorescence. A549 cells replicateplated on teflon-coated Meloy slides (I2Materials and Methods well) were grown to uniform subconfluent Cells. A549 cells (American Type Tissue layers. The medium was gently aspirated and Culture Collection, Rockville, MD) were replaced by medium supplemented with i o% grown in complete Waymouth's media sup- NU-serum containing one of the following: plemented with IO% NU-serum (Collabora- glucose oxidase (o.o i u/ml), H202 tive Research, Lexington, MA) and plated (3 X IO-5M), glucose oxidase boiled for 20 onto Falcon microtiter plates (Falcon Lab- min, boiled glucose oxidase plus catalase, H202 plus catalase, and H202 plus boiled ware, Oxnard, CA) for 51Cr release experiments, or on 12 well Meloy slides (Meloy catalase. Only one set of reagents was used Laboratories, Springfield, VA) for analysis of per slide. Duplicate, coded slides were placed in a petri dish and incubated for intervals the cytoskeleton. varying from I to 4 hours. Medium contain51Chromium Release Assay. A549 cells ing colchicine (40 jug/ml) and cytochalasin B labelled with sodium chromate (5'Cr) (New (o.s,pg/ml) were used as positive controls England Nuclear, Boston, MA) were exposed (Knapp et al. i983a, b). Untreated cells were included on every slide. One slide was used to H202 or glucose oxidase in complete Waymouth's medium (27 mM glucose) as for immunofluorescence and the other for trypan blue dye exclusion. After incubation previously described by Harlan et al. (I982; I984). Briefly, after overnight incubation of the medium was removed and the cells were 5'Cr-labelled cells, the monolayer was washed gently in phosphate buffered saline without calcium and magnesium, fixed in washed three times with I% NU-serum in absolute ethanol for 5 min and processed for phosphate buffered saline (PBS) (GIBCO, indirect immunofluorescence. A monoclonal Grand Island, NY), and incubated for IO min at 3 70C in a 5% CO2 atmosphere, with either antibody to cytokeratin (54 000 mol. wt fragment) and a polyclonal antibody against I,3-bis(chloroethyl)-I-nitrosourea (BCNU), IO0 ,ug/ml in o. I% ethanol, or o. i% ethanol actin were used as primary antibodies, prepared as previously described (Gown & Vogel in io% NU-serum in Waymouth's media. I982; Groschel-Stewart et al. 1977). DupliFollowing two additional washes with I% cate slides were incubated with o. i% trypan NU-serum in PBS, the test or control medium blue for 5 min and fixed with 2% glutaraldewas added to a final volume of IO0odl per well hyde. Percent viability was calculated as and the cells were incubated at 3 70C in a 5% number of live cells/total number of CO2 atmosphere. Supernatant medium was

H202-induced cytoskeletal changes cells x i oo. All experiments were repeated at least three times. Reagents. Glucose oxidase, H202 catalase, colchicine and cytochalasin B were obtained from Sigma Chemical Company (St Louis, MO). I, 3-bis(chloroethyl)- I -nitrosourea (BCNU) was obtained from Bristol Laboratories (Syracuse, NY), and dissolved in ethanol at a concentration of IOO ,ug/ml just before each experiment.

Results 51Cr Release and Trypan Blue Dye Exclusion 51Cr release was both time- and dose-dependent (Fig. ia and b). There was no percent specific 5'Cr release from cells exposed to concentrations of glucose oxidase varying from O.OO i u/mI to 0.05 u/ml at 4 h. Greater a

11(

110 _

107

than 90% of the cells also excluded trypan blue dye at this point. A dose-dependent, percent specific 51Cr release was demonstrated from cells exposed to varying concentrations of glucose oxidase (o.ooi u/ml to 0.05 u/ml) for 8 h (Fig. ia). Significant percent specific 51Cr release (43%) was first seen after 8 h of exposure to glucose oxidase (0.05 u/ml) (Fig. ib) or H202, at which time more than 8o% of the cells were trypan blue dye positive. At 24 h percent specific 5 Cr release was 69%. In cells pretreated with BCNU, there was greater percent specific 51Cr release at 8 h (64% vs 43%), and complete lysis occurred after 24 h of exposure to glucose oxidase or H202. The 5'Cr release was abrogated by the addition of catalase to glucose oxidase or H202, or by heat inactivation of glucose oxidase. (% specific 51Cr release =43% with glucose oxidase alone; 3% when catalase was added). b

100 _ 90 _ 80 CD

U' C)D

70 60

2

50

U)O'

40 3

,.

0.001 0.005 0.01 0.05 Glucose oxidase (units/ml) at 8 h

0

4

8

12 16 20 Time (hours)

24

Fig. i. Time and dose response of glucose-glucose oxidase-induced 51Cr release from A549 cells: cells were labeled overnight with 51Cr, washed and preincubated for ten minutes with o. i% ETOH (triangle) or BCNU I100 ig/ml (open circle). They were again washed and varying concentrations of glucose oxidase were added (a), or a fixed concentration (0.05 u/ml; b). Medium was sampled at 8 h (a) and at 4, 8, and 24 h (b), to determine specific 5'Cr release. Values represent mean ± i SE (n = 8).

Io8

G. Raghu et al.

Fig. 2. Untreated A549 cells stained with antibodies to cytokeratin and actin filaments. Prominent filamentous array of cytokeratin strands radiating from the nuclear margin to the periphery (a); fine filamentous strands of actin present throughout the cytoplasm (b). x 378 using Leitz epifluorescence microscope. [photographed using Ilford P5 (ASA 400) film]

Distribution of Cytoskeletal Filaments and Trypan Blue Dye Exclusion In untreated A549 cells cytokeratin was present as a prominent, uniform, and smooth linear filamentous network extending from the nuclear margin to the outer cell membrane (Fig. 2a). Actin was distributed in a fine, regular filamentous pattern throughout the cytoplasm, with areas of increased density at the plasma membrane (Fig. 2b). In the cells treated with glucose oxidase (o.o i u/ml) and H202 (3 x Io-5M) for 2 and 4 hours, there was disruption of cytokeratin filaments. The filaments were irregular, with gaps in their array, and appeared as 'granules' scattered throughout the cytoplasm in greater than 90% of the cells at 4 h (Fig. 3a). At 2 h the changes were less prominent and

were seen in only 30-40% of the cells exposed to glucose oxidase and H202. Actin microfilaments were partially fragmented at 2 hours and completely disrupted in greater than 90% of the cells exposed to glucose oxidase and H202 for 4 h. Those remaining were seen as a tangled array with granular aggregates at the plasma membrane (Fig. 3b). Similar changes in cytokeratin and actin filaments were seen in cells exposed to colchicine and cytochalasin B (not shown).

Discussion Oxygen metabolites may damage tissues and cause increased microvascular permeability (Del Maestro et al. i98 ia, b). The role of one metabolite, H202, has been well established

H202-induced cytoskeletal changes

lO9

NADP

Reduced glutathione 91 (2GSH)

Glutathione BCNU *-reductaseI

Glutathione peroxidase

(GR)

NADPH + H+

H0 H202

(GPO) Oxidised glutathione

2H20

(GSSG)

Fig. 4. Glutathione peroxidase-glutathione reductase system (glutathione redox cycle). I,3-bis(chloroethyl)-i)-nitrosourea (BCNU) inhibits glutathione reductase (adapted from Harlan et al. I984).

both in vivo and in vitro (Balin et a]. I976; Johnson et al. I98I; Ley & Arfors I982; Suttorp & Simon I 982; Tate et al. I 982; Weiss et al. I 98 I). In the present study, there was a time- and dose-dependent injury to A549 cells with either the direct addition of H202 or H202 generated by glucose-glucose oxidase. The addition of catalase to glucose oxidase or H202, or the inactivation of

glucose oxidase by heat, abrogated the 51Cr release, indicating that H202 was the mediator of injury. Comparable cytotoxicity by enzymatically generated H202 has been demonstrated in endothelial cell monolayers (Weiss et al. I 98 I). Glutathione reductase is one means by which intracellular H202 is metabolized (Harlan et al. I984; Nathon et al. I98I). In

iio

H202-induced cytoskeletal changes

the glutathione redox cycle (Fig. 4), the oxidized glutathione is reconverted to reduced glutathione by glutathione reductase. Glutathione reductase may be selectively inhibited by BCNU. This phenomenon has been exploited in studies of the antitumour effects of H202 (Nathan & Cohn I 98 I; Nathan et al. I98I). In the present study, pretreatment with BCNU potentiated the 51Cr release, suggesting that one facet of the H202-induced cell injury is intracytoplasmic. Cytoskeletal changes were detected earlier than 51Cr release or trypan blue dye uptake in this study. Furthermore, they occurred with sublethal doses of glucose oxidase. The earliest detectable changes, seen at 2 h, were beading or fragmentation of micro- and intermediate filaments. At 4 h these changes were more pronounced and present in greater than 90% of the cells, at which time 90% of the cells still excluded trypan blue dye. Significant 51Cr release occurred only after cytoskeletal changes, and required a five-fold greater concentration of glucose oxidase. Little is known about cytoskeletal changes in response to, or as a consequence of, cell injury, with the exception that increased endothelial permeability in vitro was associated with disrupted microfilaments (Shasby et al. I982). The integrity of tight junctions and of an intact permeability barrier in MDCK epithelial cell layers has been correlated with the presence of a normal cytoskeleton (Meza et al. I980). H202 has recently been implicated in injury to MDCK epithelial cells resulting in decreased transpithelial resistance, an indicator of ion permeability (Parsons et al. I984). Changes in cytokeratin and actin in A549 cells were an early result of oxidant injury in this study, and it is likely that similar changes occur in other cell types in vitro. A549 cells are a malignant cell line derived from human lung carcinoma. While these cells have several morphological and biochemical characteristics in common with type II pneumocytes (Leiber et al. I 9 76; Nardone & Andrews I979; Nardone et al. I982), they have

distinct differences in their lipid content compared to that of freshly isolated rat Type II pneumocytes (Mason & Williams I980). It is not known whether H202 or other toxic agents would induce similar changes in normal pulmonary epithelial cells in vitro. However, the fact that H202 mediates injury to MDCK and A549 cells in vitro suggests it may affect other epithelial cell types including alveolar epithelial cells. In the present study both micro- and intermediate filaments were altered in cells exposed to H202, but it was not determined whether the injury was directed against one or both filaments or whether the effects were somehow linked, as was shown in the mouse KLN 205 carcinoma cells and HELA cells treated with colchicine and cytochalasin D (Knapp I983a, b). The biochemical alterations which underline disruption of cytoskeletal filaments in H202-induced cell injury have not been elucidated. Oxidant agents may mediate cell injury through lipid peroxidation of the cell membrane (Riely et al., I9 74), or protein denaturation by formation of disulfide bonds (Badwey & Karnovsky I980). The cytoskeletal changes reported here may therefore have resulted from either some direct effect of H202 on the cytoskeleton or from an indirect effect. The exact nature of the injury awaits further study.

Acknowledgements This research was supported by HL30542 and HLi 8645 from the National Institutes of Health, Bethesda, MD. Dr Raghu is the recipient of a National Research Service Award, HLo6745 from the National Institutes of Health, Bethesda, MD. The technical assistance of James Hibbert and manuscript preparation by Bev Wyrick are gratefully acknowledged. References BADWEY J.A. & KARNOVSKY M.L. (I980) Active oxygen species and the functions of phagocytic leukocytes. Ann. Rev. Biochem. 49, 695-726. BALIN A.T., GOODMAN D.B.P., RASMUSSEN H. &

H202-induced cytoskeletal changes CRISTOFALO V.J. (I 9 76) The effect of oxygen tension on the growth and metabolism of WI38 cells. J. Cell. Physiol. 89, 235-249. DEL MAESTRO R.F., BJORK J. & ARFORS K.E. (198 Ia) Increase in microvascular permeability induced by enzymatically generated free radicals. I. In vivo study. Microvasc. Res. 22, 239254. DEL MAESTRo R.F., BJORK J. & ARFos K.E. (I98 ib) Increase in microvascular permeability induced by enzymatically generated free radicals. II. Role of superoxide anion radical, hydrogen peroxide, and hydroxyl radical. Microvasc. Res. 22, 255-270. GOWN A.M. & VOGEL A.M. (i982) Monoclonal antibodies to intermediate filament proteins of human cells-unique and crossreacting antibodies. J. Cell Biol. 95, 4 I4-424. GROSCHEL-STEWART U., CEURREMANS S., LEHR I., MAHLMEISTER C. & PAAR E. (I977) Production of specific antibodies to contractile proteins and their use in immunofluorescence microscopy. Histochem. 50, 271-279. HARLAN J.M., KILLEN P.D., HARKER L.A., STRIKER G.E. & WRIGHT D.G. (I982) Neutrophil mediated endothelial injury in vitro: mechanisms of cell detachment. 1. Clin. Invest. 68, I394-I403.

HARLAN J.M., LEVINE J.D., CALLAHAN K.S.,

SCHWARTZ B.R. & HARKER L.A. (I984) Glutathione redox cycle protects cultured endothelial cells against lysis by extracellularly-generated hydrogen peroxide. J. Clin. Invest. 73, 706-71 3. JOHNSON K.J., FANTONE J.C., KAPLAN J. & WARD P.A. (I98I) In vivo damage of rat lungs by oxygen metabolites. J. Clin. Invest. 67, 983993. KNAPP L.W., O'GuIN W.M. & SAWYER R.H. (i983a) Drug induced alterations of cytokeratin organization in cultured epithelial cells. Science 219, 501-503. KNAPP L.W., O0GUIN W.M. & SAWYER R.H. (i 983b) Rearrangement of the keratin cytoskeleton after combined treatment with microtubule and microfilament inhibitors. J. Cell Biol. 97, I788-I 794. LAZARIDES E. (I982) Intermediate filaments: a chemically heterogenous, developmentally regulated class of protein. Ann. Rev. Biochem. 51,2I9-250. LAZARIDES E. (I980) Intermediate filaments as mechanical integrators of cellular space. Nature 283, 249-256. LEIBER M., SMITH B., SZAKAL A., NELSON-REES W. & TODARO G. (I976) A continuous human cell line from a human lung carcinoma with pro-

III

perties of Type II alveolar epithelial cells. Int. 1. Cancer 17, 62-70. LEY K. & ARFORS K.E. (I982) Changes in macro-

molecular permeability by intravascular generation of oxygen-derived free radicals. Microvasc. Res. 24, 25-33. MARTIN W.J., GADEK J.E., HUNNINGHAKE G.W. &

CRYSTAL R.G. (I982) Oxidant injury of lung parenchymal cells. J. Clin. Invest. 68, 1277-

1288. MASON R.J. & WILLIAMS M.C. (I980) Phospholipid composition and ultra-structure of A549 cells

and other cultured pulmonary epithelial cells of presumed Type II cell origin. Biochim. Biophys. Acta 617, 36-50. MEISTER A. (I983) Selective modification of glutathione metabolism. Science 220, 472-477. MEZA I., IBANA G., SABANEVO M., MATINEZPALOMA A & CEVERGIDO M. Occluding junctions and cytoskeletal components in a cultured transporting epithelium. J. Cell Biol. 87, 746754. NARDONE L.L. & ANDREWS S.B. (I979) Cell line A549 as a model of the Type II pneumocytephospholipid biosynthesis from native and organometallic precursors. Biochim. Biophys. Acta 573, 276-295. NARDONE L.L., POCHRON S. & SZOKE E. (1 982) Effect of lung structural matrix on expression of differentiated properties of Type II alveolar cells. 1. Cell Biol. 95, 134. NATHAN C.F. & COHN Z.A. (I98I) Antitumor effects of hydrogen peroxide in vivo. 1. Exp. Med. 154, 1539-1553.

NATHAN C.F., ARRICK B.A., MIJRRAY H.W., DESANTIS N.M. & COHN Z.A. (I98I) Tumour cell antioxidant defenses: Inhibition of glutathione redox cycle enhances macrophage mediated cytolysis. J. Exp. Med. 1I53, 766-782. PARSONS P.E., Corr G.R., MASON R.J. & HENSON P.M. (I984) Reversible oxidant injury of an epithelial cell monolayer. Fed. Proc. 43, 777. RIELY C.A., COHEN G. & LIEBERMAN M. (I 9 74) Ethane evolution: a new index of lipid peroxidation. Science I83, 208-210. Roos D., WEENING R.S., VOETMAN A.A., VAN

SCHAIK M.L., BOT A.A., MEERHORF L.J. & Loos J.A. (I979) Protection of phagocytic leukocytes by endogenous glutathione: Studies in a family with glutathione reductase deficiency. Blood 53, 85i-866.

SACKS T., MOLDOW C.F., CRADDOCK P.R., BOWERS T.K. & JACOBs H.S. (1978) Oxygen radicals mediate endothelial cell damage by comple-

ment-stimulated granulocytes. J. Clin. Invest. 6i, II6I-II67.

II2

G. Raghu et al.

SCHLIWA M. & VAN BLERKOM J.V. (I98I) Structural interaction of cytoskeletal components. 1. Cell Biol. 90, 222-235. SHASBY D.M., SHASBY S.S., SULLIVAN J.M. & PEACH M.J. (i 982) Role of endothelial cell cytoskeleton in control of endothelial permeability. Circ. Res. 5I, 657-66I. SurroRP N. & SIMON L.L. (I 9 8 2) Lung cell oxidant injury-enhancement of polymorphonuclear leukocyte mediated cytotoxicity in lung cells exposed to sustained in vitro hyperoxia. 1. Clin. Invest. 70, 342-350. TATE R.M., SHASBY D.M., VAN BENTHUYSEWN K.M., MCMURTY I.F. & REPINE J.E. (I982) Oxygen

radical-induced pulmonary edema. Chest 8I, 5 7s-5 9s. WEBER K. & OSBORN M. (I982) The cytoskeleton. Nati. Cancer Inst. Monogr. 63, 3I-46. WEISS S.J., YOUNG J. LoBUGLIo A.F., SLIVKA A. & NEMEH N.F. (I98I) Role of hydrogen peroxide in neutrophil-mediated destruction of cultured endothelial cells. J. Clin. Invest. 68, 7I4-72I. YAMADA 0., MOLDOW C.F., SACKS T., CRADDOCK P.R., BOOGAERTS M.A. & JACOBS H.S. (I98I) Deleterious effects of endotoxin on cultured endothelial cells: An in vitro model of vascular injury. Inflammation 5, I I 5-I 26.