Cell Motility and the Cytoskeleton 16:llO-120 (1990)
Cytoskeletal Dynamics in Rabbit Synovial Fibroblasts: 1. Effects of Acrylamide on Inte rmediate Filaments and Microfilaments Judith Aggeler and Keith Seely
Department of Human Anatomy, School of Medicine, University of California, Davis, California Rabbit synovial fibroblasts respond to changes in cell shape and cytoskeletal architecture by altering specific gene expression. We have tested the ability of acrylamide, a neurotoxin that alters the distribution of intermediate filaments in cultured PtKl cells, to induce metalloprotease expression in synovial fibroblasts. Cells treated with 2-20 mM acrylamide for 5 to 24 h underwent shape changes similar to cells treated with the tumor promoter phorbol myristate acetate. Intermediate filaments visualized with anti-vimentin antibodies did not collapse into a perinuclear cap in these rounded cells, but were still present in the extended cell processes. Unexpectedly, when actin was visualized in acrylamide-treated cells, extensive dissociation and clumping of microfilaments was observed. Concentrations of acrylamide > 10 mM were cytotoxic, but cells recovered completely after 24 h incubation with 5 mM acrylamide. Like other agents that alter cell shape and actin distribution in synovial fibroblasts, acrylamide also induced expression of the secreted metalloprotease collagenase. Although some recent evidence suggests that acrylamide may be able to exert its collagenase-inducing effects extracellularly , perhaps through transmembrane matrix receptors, our observation that this neurotoxin dramatically alters protein synthesis in synovial fibroblasts suggests that direct effects on cell metabolism may also play a role in acute acrylamide intoxication. Key words: vimentin, collagenase, cell shape, gene expression
INTRODUCTION The cytoskeleton of mammalian cells is made up of three major filament systems: microfilaments, intermediate filaments, and microtubules. A vast literature exists documenting the varied structure, function, and dynamics of actin filaments and microtubules. Although the underlying structure of intermediate filaments has recently been described in some detail [Steinert et al., 19851, the functions of this filament system have not yet been fully elucidated. This anomalous situation has arisen in part because the intermediate filament networks of most cells are relatively stable, making studies of filament assembly and turnover difficult, and in part because, unlike other cytoskeletal components, no agent has previously been available that can perturb intermediate filaments with reasonable specificity. Recent studm innn iaxi,..,
T
:-- T.,.
ies by Eckert [1985, 19861, have shown that the neurotoxin acrylamide can collapse the intermediate filament network of cultured PtK 1 cells without obvious effect on actin or microtubules. Such activities as vesicular movement and membrane ruffling continued in acrylamidetreated cells, indicating that intermediate filaments are not central to the mechanism(s) underlying these vital
Received June 22. 1989: accepted February 1. 1990. Address reprint requests to Dr. Judith Aggeler, Department o f Human Anatomy. University of California, Davis, CA 95616. Abbreviations used: DME. Dulhecco’s modification of Eagle‘s minimum essential medium: HBSS. Hanks’ balanced salt solution: PMA, phorhol myristate acetate: SDS-PAGE, sodium dodecylsulfate polyacrylamide gel electrophoresis.
Actin Dynamics in Acrylamide-Treated Cells
111
cell functions. These observations form the basis for further experimental investigation into the role of intermediate filaments in many cell functions that are thought to depend, at least in part. on components of the cellular cytoskeleton. We have been interested for a number of years in the mechanism(s) leading to the induction and sustained expression of metalloproteases (collagenase and stromelysin) in cultured fibroblasts [Werb and Aggeler, 1978; Aggeler et al., 1984a,b; Chin et al., 1985; Werb et al., 19861. A variety of agents with pharmacologic activities implicating plasma membrane receptors [Werb and Aggeler, 1978; Aggeler et al., 1984b; Werb et al., 19891, intracellular Ca mobilization [Aggeler et al., 1984b; Unemori and Werb, 19881, and protein kinases [Aggeler et al., 1984al have been identified as inducers. An important correlate to these studies has been the observation that induction of metalloprotease expression in cultured cells is preceded by changes in cell morphology (cell rounding) [Werb and Aggeler, 1978; Aggeler et al., 1984bl. In addition, it is clear that expression of collagenase and stromelysin is induced by a relatively short exposure to most pharmacological agents (1-6 h) and that once induced, synthesis and secretion can continue for many days in the absence of the agent [Aggeler et al., 1984bl. Present evidence strongly suggests that the cell shape changes associated with collagenase induction are based on alterations of the actin cytoskeleton [Aggeler et al., 1984b; Unemori and Werb, 19861 and that microtubules are not directly involved [Unemori and Werb, 19861. Intermediate filaments have not previously been implicated in this modulation of gene expression; however, the recent report of Green and colleagues 119861 that microfilaments and intermediate filaments interact during spreading of rounded cells in culture suggested to us that this filament system might also be involved in other cell functions previously thought to be solely actinbased. We now report our attempts to determine the role played by intermediate filaments in metalloprotease induction by treating cultured rabbit synovial fibroblasts with acrylamide. The accompanying paper [Aggeler, 19901 documents the redistribution of microfilaments and intermediate filaments during the period of sustained +
+
Fig. I . Acrylamide-induced shape changes in rabbit synovial fibroblasts. Rabbit synovial fibroblasts were incubated in the absence (a) or presence ofacrylamide (S m M ) for 6 (b), 8 ( c ) .or 16 h ( d ) .With time. acrylamide-treated cells retract from each other and become increasingly rounded. Open arrows indicate retraction fibers. These morphological changes art similar t o those observed in PMA-treated cells, where the degree o f shape change has been shown to predict the extent of induced collagenase synthesis (see Aggeler et al., 1984a. Fig. I ) . x 300.
112
Aggeler and Seely TABLE I. Acrylamide Inhibits Protein Synthesis* Protein synthesis Cells Treatment Control Acrylamide
Medium
Time (h)
Incorporation (cpm x lo4)
Inhibition
lncorporation (cpm x lo4)
Inhibition
(or,)
18
450 268 129
40.4 71.3
2.5 2.8 2.9
-
6 18
(%)
0 0
*Rabbit synovial fibroblasts incubated in serum-free medium in the absence or presence of 5 mM acrylamide for the times indicated were labeled for 2 h with ["SI-methionine. Total protein synthesis was determined by counting an aliquot of the detergent-solubilized cell layer. Biosynthesis of secreted protein was determined by precipitating from the medium with trichloroacetic acid and counting. Inhibition of protein synthesis by acrylamide is expressed as a percentage of the untreated control cell incorporation.
and maintained in Dulbecco's modified Eagle's medium (DME) supplemented with 10% fetal bovine serum (both C 6 2 4 4 8 from GIBCO, Grand Island, NY), as described previously [Aggeler et al., 1984al. Cells were subcultured weekly ( 1 5 ) with 0.25% trypsin and used between passages 2 and 10. Cells were plated onto coverslips (1 2 mm dia.) in 12-well plates (Costar) in DME plus serum 16 to 24 h before washing with Hanks' balanced salt solution (HBSS) and addition of serum-free DME supplemented with 0.2% lactalbumin hydrolysate. For most experiments cells were treated with acrylamide (electrophoresis grade; BioRad, Richmond, CA) in serum-free medium CL containing 0.2% lactalbumin hydrolysate for the times indicated. Serum-free medium has been used in this experimental system in the past to facilitate detection of collagenase activity. The morphological changes observed in synovial fibroblasts in the presence of acrylamide (see below) are also observed in cells treated in the presence of serum; however, recovery from treatment 1 2 3 4 after toxin removal is retarded in the absence of serum Fig. 2. Effects of acrylamide on synthesis of secreted proteins. Rabbit [Aggeler, 19901. Concentrations of acrylamide 5 5 mM synovial fibroblasts were incubated for 6 h in the absence (lane 1) or were not toxic to synovial fibroblasts, as judged by trypresence (lane 2-4) of 5 mM acrylamide, at which time they were pan blue exclusion (>95% of cells negative)(not shown) either radiolabeled immediately (lanes 1.2) or allowed to recover from acrylamide treatment in fresh serum-free medium for an additional 18 and by the rapid respreading of the cells after acrylamide (lane 3) or 42 h (lane 4)(total incubation 6, 24, or 48 h). This brief was removed [Aggeler, 19901. In come control experitreatment with acrylamide induces sustained procollagenase gene ex- ments, cells were treated with phorbol myristate acetate pression (CL) in respread cells after 24-48 h (lanes 3.4). The secreted (PMA)(20 nglml) or colchicine ( lop6 M)(Sigma Chempolypeptide migrating at M, = 53,000 has previously been identified ical Corp., St. Louis, MO), as described for acrylamide as procollagenase by immunoprecipitation and peptide mapping above. PMA was dissolved in ethanol at 1 mg/ml and [Aggeler, et al., 1984aj. stored at -20°C. Colchicine was dissolved in water at lo-' M and stored at 4°C. collagenase synthesis and secretion that follows withdrawal of inducing agents. Labeling of Proteins and Sodium Dodecylsulfate-PolyacrylamideGel MATERIALS AND METHODS Electrophoresis (SDS-PAGE) Cell Culture Cells plated into 12-well culture dishes (25 mm dia., Costar) and exposed to CB, CD, or acrylamide Rabbit synovial fibroblasts were harvested from the were pulse-labeled with 50 pCi/ml [%I-methionine knee joints of 4-6-week old New Zealand white rabbits x
-
Actin Dynamics in Acrylamide-Treated Cells
113
Fig. 3. Effects of acrylamide, PMA, and colchicine on intermediate filaments. Rabbit synovial fibroblasts were treated with acrylamide ( 5 mM) (a), PMA (20 ngiml) (b). or colchicine (lo-' M ) ( c ) for 16 h before fixation and staining for intermediate filaments (vimentin). Note that intermediate filament bundles appear to fill the numerous
retraction fibers present in both acrylamide- and PMA-treated cultures (open arrows). In contrast, colchicine (c) causes partial collapse of the intermediate filament network away from the cell periphery (arrowheads) into a perinuclear location. Unlike acrylamide and PMA, colchicine does not cause cell rounding o r retraction. X 300.
(1,350 Ci/mmol; Amersham Corp., Lisle, IL) in methionine-free DME. After 2 h, radiolabeled secreted protein was precipitated from the medium by the addition of cold trichloroacetic acid (final concentration 5 % ) . The washed protein pellets were resuspended in Laemmli sample buffer containing 0.5% P-mercapto-
ethanol and boiled for 3 min before electrophoresis [Laemmli, 19701. SDS-PAGE was carried out using 12% minigels with a 3% stacking gel. Dried gels were autoradiographed with Kodak X-Omat AR x-ray film, preflashed before exposure, as described previously [Aggeler et al., 1984al.
114
Aggeler and Seely
Fig. 4. Microfilament disassembly in acrylamide-treated cells. Rabbit synovial fibroblasts were incubated in serum-free medium in the absence (a)or presence ofacrylamide ( 5 mM) for S ( b ) or 18 h ( c ) before fixation and staining o f actin with rhodamine-phalloidin. At early times (b), a few treated cells have begun to retract. and some clumping of actin and decreased numbers of stress fibers are observed. even in
apparently spread cells (arrows). By 18 h (c) a majority of the cells have rounded to some extent and virtually all cells in the culture show some degree of actin clumping Some of the larger actin clumps are clearly visible in the accompanying phase contrast images of the same cells (arrowheads). X 300.
Actin Dynamics in Acrylamide-Treated Cells
115
Fig. 5 . Actin distribution in acrylamide-treated cells. Rabbit synovial fibroblasts were incubated in the absence (a) or presence (b) of acrylamide ( 5 mM) for 18 h before fixation and staining of actin with a monoclonal anti-actin antibody. Disassembly and clumping of microfilaments in acrylamide-treated cells is similar to that observed with rhodamine-phalloidin (Fig. 4).A moderate degree of diffuse cytoplas-
mic staining in both controls (a) and treated cells (b) indicates the presence of G-actin, which does not appear to be altered in acrylamide-treated cells. As in the rhodamine-phalloidin-stained cells (Fig. 4), the actin clumps are also visible in the phase contrast images of the same cells (arrowheads, b). X 300.
lmmunofluorescence Microscopy
sham). Intermediate filaments were visualized with a mouse monoclonal antibody raised against chicken gizzard vimentin (ICN Immunobiologicals, Lisle, IL), followed by biotinylated rabbit anti-mouse IgG (Amersham) and streptavidin-fluorescein (BRL, Bethesda, MD) or streptavidin-Texas Red. Cells were observed under epiilumination using a Zeiss Photomicroscope 111 equipped with a Plan NEOFLUAR 25 objective, N.A. 0.8, and photographed with Kodak Tri-X 400 film.
Fibroblasts cultured on 12 mm diameter glass coverslips were incubated with inducing agents for 1 to 24 h. Cells were then washed and fixed with freshly prepared 2% paraformaldehyde in phosphate buffered saline, pH 6.8, for 15 min at 22°C. Fixed cells were permeabilized by immersing them for 5 min in cold acetone (-20°C). Actin filaments were visualized with rhodamine-phalloidin (30 min, 22"C), as described by the manufacturer (Molecular Probes, Junction City, OR). In some experiments, actin was detected using a monoclonal antibody directed against chicken gizzard actin (Amersham Corp., Arlington Heights, IL), followed by biotinylated rabbit anti-mouse IgM and streptavidin-Texas Red (Amer-
Transmission Electron Microscopy
For transmission electron microscopy, rabbit synovial fibroblasts plated in 16 mm culture wells were incubated with 5 mM acrylamide in serum-free medium for
116
Aggeler and Seely
20 h. Cells were fixed with 2% glutaraldehyde (in Hanks' balanced salt solution containing 20 mM HEPES buffer, pH 6.8) for 1 hr at room temperature, postfixed with 0.5% OsO, for 15 min at 4"C, dehydrated, and embedded en bloc in EpodAraldite. Pieces of hardened Epon were then reembedded so that thin sections cut perpendicular to the dish surface could be prepared. Sections were observed using a Philips 400 electron microscope. RESULTS Alteration of Cell Shape in Rabbit Fibroblasts Treated With Acrylamide
Rabbit synovial fibroblasts treated with 2-20 mM acrylamide for up to 24 h showed dose-dependent changes in cell shape. This cell rounding was progressive with time, as illustrated in Figure 1 for cells treated with 5 mM acrylamide. After 5-6 h exposure, a few cells were rounded and showed spindly retraction fibers (Fig. lb). By 8 h, cell rounding and retraction were much more pronounced (Fig. lc), and after 16 h, virtually all cells in the culture showed morphological changes (Fig. Id). Essentially the same result was obtained for cells treated in the presence or absence of serum. Concentrations of acrylamide 95%) and by respreading and growth of cells after the drug was removed [Aggeler, 19901. Higher concentrations (10-20 mM) routinely killed cells, even after relatively brief exposures ( 5 h). This effective concentration range is similar to that observed previously for other cultured cells [Eckert, 19851. Acrylamide Induces Procollagenase Synthesis and Secretion
We have previously shown that a variety of agents that cause cell shape changes in rabbit synovial fibro-
Fig. 6. Thin section views of acrylamide-induced actin clumps. Rabbit synovial fibroblasts were incubated with 5 mM acrylamide for 20 h, fixed, and embedded prior to sectioning. A low magnification view of an entire cell (a) reveals that large aggregates of actin are located in the perinuclear cytoplasm (arrow), well above the basal surface o f the cell. At high magnification (b), this aggregate is seen to consist of numerous small dense clumps of thin filaments (open arrows). Cell organelles appear to be excluded from these aggregates, and even intermediate filaments. numerous in the surrounding cytoplasm (closed arrow). are not observed within these areas. In (c). an actin aggregate at the basal surface (arrowheads) of a cell treated with cytochalasin D for 20 h is shown for comparison. m = mitochondrion. a: Bar = 2 p m , X6.900. b.c: Bar = 0.2 pm, X90.000.
Actin Dynamics in Acrylamide-Treated Cells
blasts cause these cells to undergo a specific switch in gene expression characterized by down regulation of collagen and fibronectin synthesis and induction of the extracellular matrix-degrading metalloproteases collagenase and stromelysin [Werb and Aggeler, 1978; Aggeler et al., 1984a,b; Unemori and Werb, 19861. When newly synthesized secreted proteins of acrylamide-treated fibroblasts were examined by SDS-PAGE, a similar induction of collagenase expression was observed (Fig. 2). Because protein synthesis was strongly inhibited (40-70% reduction in ["S]-methionine incorporation into cellular protein) in the presence of acrylamide (Table I), induction of collagenase was most clearly shown in cells that had respread after removal of the drug.
Intermediate Filament Distribution in Acrylamide-Treated Cells
Because cultured PtKl cells exposed to S mM acrylamide show dramatic alterations in the distribution of the intermediate filament cytoskeleton (both vimentin and cytokeratin)[Eckert, 19851, and because both cell shape changes and collagenase induction were observed in acrylamide-treated synovial fibroblasts, we were particularly interested in determining whether the intermediate filaments of these cells were also affected by the drug. When the intermediate filament mesh of acrylamide-treated rabbit synovial fibroblasts was observed by immunofluorescence microscopy using an anti-vimentin antibody, wholesale collapse and retraction of intermediate filaments was not observed (Fig. 3a), although some aggregation of individual filaments into bundles may occur. These bright-staining bundles of vimentin filaments were present throughout the cytoplasm of rounded cells and extended out into the numerous retraction fibers. This pattern of intermediate filament distribution was qualitatively similar to that observed in cells treated with another collagenase-inducing agent, the phorbol ester tumor promoter, PMA (Fig. 3b), as described previously by Laszlo and Bissell [1983]. In contrast, the microtubule-depolymerizing alkaloid, colchicine, neither alters the overall morphology or microfilament distribution of rabbit synovial fibroblasts [Werb and Aggeler, 1978; Unemori and Werb, 19861 nor induces collagenase [Unemori and Werb, 19861; however, it does cause collapse of intermediate filaments toward the nucleus (Fig. 3c), as has been described previously for a variety of cultured cells [Franke et al., 19781. These results indicated that the dramatic changes in cell shape observed in acrylamide-treated fibroblasts were not correlated with distinctive changes in intermediate filament distribution.
117
TABLE 11. Acrylamide-Induced Changes in Actin Cytoskeleton* Morphology Treatment ~~
Time (h) Unchanged
Control Acrylamide (mM) 2
5
24
97
3
0
5 18 24
95 33 12
4 48 36
19 52
5
63 0 0
32 0 I1
4 100 89
18
24 ~~
Intermediate ClumDed
~
1
~
"Rabbit synovial fibroblasts treated with acrylamide for the times and at the concentrations indicated were fixed and stained with rhodamine phalloidin to visualize actin filaments. One hundred randomly-chosen cells were scored at the microscope for each treatment; values represent numbers of cells scored in each category. Moderate changes in cell morphology and actin cytoskeleton were characterized by some cytoplasmic retraction and condensation of filamentous actin (for example, see Fig. 4b). Rounded cells showed no sign of stress fibers and were characterized by localization of actin in several large cytoplasmic clumps.
Alteration of Actin Cytoskeletal Architecture in Rabbit Fibroblasts Treated With Acrylamide
Having found no major alterations of the intermediate filament mesh of acrylamide-treated fibroblasts, we next sought to determine the effects of acrylamide on the actin cytoskeleton of these cells. In contrast to intermediate filaments, the pattern of actin filament distribution was dramatically altered in cells rounded by exposure to acrylamide (Fig. 4). In parallel to the changes in cell shape observed in Figure 1, acrylamide-treated cells showed progressive condensation and clumping of filamentous actin, accompanied by loss of distinctive microfilament bundles or stress fibers. With time, increasing numbers of cells in these cultures showed these changes in actin distribution (Table II), and by 16-18 h of treatment with S mM acrylamide, virtually every cell displayed some degree of actin redistribution. In most cells, actin was completely collapsed into a number of bright-staining aggregates (Fig. 4c). When acrylamidetreated cells were stained using an anti-actin antibody, the same pattern of actin clumping was observed (Fig. 5b). Cells also displayed a moderate level of diffuse cytoplasmic actin-staining (G-actin), which appeared to be somewhat reduced by treatment with acrylamide (Fig. Sb). Comparison of the phase and fluorescence images of the same cells shows that the actin clumps are clearly visible by phase optics (arrowheads, Figs. 4c and 5b). Interestingly, staining of these clumps with the anti-actin antibody often appeared restricted to their surface (arrows, Fig. 5b), as if the antibody did not penetrate the aggregate, whereas rhodamine-phalloidin staining usually appeared uniform throughout the clumps.
118
Aggeler and Seely
Unlike the basally located microfilament foci induced by cytochalasin treatment (Fig. 6c) [Schliwa, 19821, the actin clumps induced by acrylamide appeared to be located in the apical cytoplasm of rounded cells, well above the plane of focus of the culture dish. This impression was confirmed by transmission electron microscopy (Fig. 6a). High magnification views indicated that these foci consisted of loose aggregates of small, dense actin clumps (Fig. 6b). Numerous intermediate filaments were found in the cytoplasm surrounding these aggregates but did not appear to be present within the clumps themselves. DISCUSSION
Although the structure and distribution of intermediate filaments have been described in some detail, their function(s) in cells have remained unclear. We have attempted to use the neurotoxin acrylamide, which alters the intermediate filament distribution of certain cultured cell types, to investigate the role played by these filaments in rabbit synovial fibroblasts. Unlike the response described for PtKl [Eckert, 1985, 19861 and PtK2 cells [Sager, 19891, the vimentin meshwork of synovial fibroblasts was not specifically affected by acrylamide. In addition, except under conditions where toxic concentrations of drug were employed ( > l o mM), we have not observed collapse of the cytokeratin filament mesh of cultured primary mammary epithelial cells treated with acrylamide either (Seely and Aggeler, in preparation). At present, we do not know why the intermediate filament networks of fibroblasts and epithelial cells in our culture systems display this stability. Both of these models use primary cells, rather than cell lines, and it may be that the transformation event(s) necessary to produce immortal lines such as PtKl alter the cytoskeleton in subtle ways. Alternatively, Eckert and Yeagle [Eckert and Yeagle, 19881 have recently shown that acrylamide causes dephosphorylation of keratins in PtKl cells, and that intermediate filament collapse can be prevented in these cells by raising intracellular levels of cyclic AMP, thus, presumably, stimulating endogenous CAMP-dependent kinase activity. Experiments are in progress to determine whether intermediate filaments in our primary cells are protected from collapse by naturally high levels of intracellular CAMP under our culture conditions. One reason for our particular interest in the intermediate filament network of synovial fibroblasts was the possibility that microfilaments and intermediate filaments might be interacting to regulate metalloprotease gene expression in these cells. Green and colleagues [Green et al., 19861 have reported that microfilaments and intermediate filaments reform from common foci in cultured cells respreading after rounding by trypsin and
cytochalasin, two agents that also induce collagenase gene expression in cultured synovial fibroblasts [Werb and Aggeler, 1978; Aggeler et al., 1984b; Harris et al., 19751. Nevertheless, as shown in the accompanying report [Aggeler, 19901, no indication of such cytoskeletal interaction was observed in synovial fibroblasts respreading after treatment with trypsin, cytochalasins B or D, or acrylamide. Taken together, our present results provide little indication for a distinct role for intermediate filaments in metalloprotease icduction and expression in cultured rabbit synovial fibroblasts and serve to reinforce the importance of the actin cytoskeleton in these processes. In contrast to the apparent lack of effect on intermediate filaments, acrylamide caused a dramatic alteration of the actin cytoskeleton of cultured rabbit synovial fibroblasts. Similar, but less dramatic, alterations in stress fibers have been noted in cultured human skin fibroblasts treated with 2 mM acrylamide for 7 days, but not in cells treated with another neurotoxin, 2,5-hexanedione [Durham et al., 19831. More recently, Sager [1989] has shown both reduction of stress fibers and actin clumping in PtK2 cells treated with acrylamide or 2,5-hexanedione, but these changes were not accompanied by cell rounding. In the present study, marked cell rounding was accompanied by disappearance of stress fibers and collapse of microfilaments into perinuclear clumps. These loose aggregates of actin were unlike those caused by cytochalasin treatment, which are typically found subadjacent to the basal plasma membrane [Schliwa, 19821; nor did they resemble the dense clumps observed by Green et al. [1986] in trypsin-treated chick embryo fibroblasts, in that intermediate filaments were not intimately associated with them. The molecular basis for the collapse of the actin cytoskeleton in acrylamidetreated synovial fibroblasts is not known, but one intriguing possibility is that this drug is acting on plasma membrane receptors for extracellular adhesion proteins, such as fibronectin or vitronectin. Eckert and Yeagle 119871 have presented preliminary data indicating that acrylamide may not enter cells. Acrylamide is known to be a very reactive molecule [Tilson, 198 11, capable of interacting with sulfhydryl groups, such as those present in the extracellular domains of the integrins [Hynes, 1987; Ruoslahti and Pierschbacher, 19871, and of cross-linking proteins [Lapadula et al., 19861. Perhaps acrlamide can bind to and alter the conformation of cell surface receptors in such a way that associated cytoplasmic anchors for actin are disrupted [Damsky et al., 1979; Burridge et al., 19881. Such a mode of action would be consistent with our observation that acrylamide-treated cells respread more rapidly in the presence of a source of exogenous adhesion molecules, such as serum [Aggeler, 19901. In addition, recent evidence indicates that colla-
Actin Dynamics in Acrylamide-Treated Cells
genase can be induced in synovial fibroblasts by antibodies directed against cell-surface fibronectin receptors [Werb et al., 19891, another indication that acrylamide may be interacting with this transmembrane signal transduction system. The effects of acrylamide on overall protein synthetic rates in synovial fibroblasts are less easy to reconcile with an extracellular site of action, although inhibition of synthesis may simply be a consequence of the profound alteration of the actin cytoskeleton, as has been observed previously in some cytochalasin-treated cells [Ornelles et al., 1986; Blum and Wicha, 19881 (Seely and Aggeler, in preparation). In these cases, it has been suggested that translation is impeded because active polyribosomes are detached from the cytoskeleton and destabilized. Alternatively, inhibition of protein synthesis in acrylamide-treated fibroblasts may be due to alterations in the state of phosphorylation of components of the synthetic machinery. Several recent studies [DeBenedetti and Baglioni, 1986; Duncan et al., 1987; Duncan and Hershey, 19891 have suggested that phosphorylation of initiation factor(s) by a cGMP-dependent kinase activity [Proud, 19861 may contribute to the inhibition of normal protein synthesis observed in heat-shocked cells. Interestingly, collagenase itself is induced by heat shock [Schorpp et al., 1984; Vance et al., 19891. Thus, acute acrylamide intoxication may invoke some aspects of a more generalized stress response in rabbit synovial fibroblasts [Angel et al., 1986; Schonthal et al., 19881. Whatever the basis for the acute inhibition of fibroblast protein synthesis by acrylamide, it is rapidly reversible. Moreover, although overall protein synthesis can be dramatically reduced, synthesis of secreted proteins continues unchanged and induction of collagenase gene expression proceeds. Thus although acrylamide has not provided the hoped-for tool with which to identify a role for intermediate filaments in metalloprotease induction, it has presented us with an additional means for studying both induction of specific gene expression and mechanisms controlling synthesis of secreted proteins in these cells. In addition, the varied responses of microfilaments and intermediate filaments of different cell types to these neurotoxins opens fertile ground for testing the possible role of factors such as extracellular matrix and phosphorylation in mediating cytoskeletal function. ACKNOWLEDGMENTS
The authors would like to thank Dr. Fern Tablin, University of California, Davis, California, for many helpful discussions and for critical reading of this manuscript. Electron micrographs were produced in the Electron Microscopy Laboratory of the School of Medicine, University of California, Davis, California. This work
119
was supported in part by a grant from the University of California Cancer Research Coordinating Committee and by NIH Biomedical Research Support Grant S07-RR05684. Some of these results have appeared previously in abstract form [Aggeler and Seely, 19871. REFERENCES Aggeler, J . ( 1990): Cytoskeletal dynamics in rabbit synovial fibroblasts. 11. Reformation of stress fibers in cells rounded by treatment with collagenase-inducing agents. Cell Motil. Cytoskel. 16: I2 1- 132. Aggeler. J . , Frisch, S . M . . and Werb, Z. (1984a): Collagenase is a major gene product of' induced rabbit synovial fibroblasts. J . Cell Biol. 98:1656-1661. Aggeler. J.. Frisch. S.M., and Werb, Z. (1984b): Changes in cell shape correlate with collagenase gene expression in rabbit synovial fibroblasts. J . Cell Biol. 98:1662-1671. Aggeler. J., and Seely, K. (1987): Effects of acrylamide on the cytoskeletons of cultured fibroblasts and epithelial cells. J . Cell Biol. 105:208a. Angel. P., Poting, A. Mallick. U., Rahmsdorf, H.J., Schorpp, M., and Herrlich, P. (1986): Induction of metallothionein and other mRNA species by carcinogens and tumor promoters in primary human skin fibroblasts. Mol. Cell Biol. 6: 1760-1766. Blum, J.L., and Wicha, M.S. (1988): Role of the cytoskeleton in laminin induced mammary gene expression. J . Cell Physiol. 135: 13-22. Burridge, K., Fath, K., Kelly, T., Nucholls, G . , and Turner, C . ( 1988): Focal adhesions: transmembrane junctions between the extracellular matrix and the eytoskeleton. Ann. Rev. Cell Biol. 4:487-525. Chin, J.R., Murphy, G., and Werb, Z. (1985): Stromelysin. A connective tissue-degrading metalloendopeptidase secreted by stimulated rabbit synovial fibroblasts in parallel with collagenase. J. Biol. Chem. 260:12367-12376. Damsky, C . H . , Wylie, D.E., and Buck, C.A. (1979): Studies on the function of cell surface glycoproteins. 11. Possible role of surface glycoproteins in the control of cytoskeletal organization and surface morphology. J . Cell Biol. 80:403-415. DeBenedetti, A , , and Baglioni, C . (1986): Activation of hemin-regulated initiation factor-2 kinase in heat-shocked HeLa cells. J. Cell Biol. 261:338-342. Duncan, R.F., and Hershey. J.W.B. (1989): Protein synthesis and protein phosphorylation during heat stress, recovery, and adaptation. J. Cell Biol. 109:1467-1481. Duncan, R . , Milburn, S . C . , and Hershey, J.W.B. (1987): Regulated phosphorylation and low abundance of HeLa initiation factor elF-4F suggest a role in translational control: heat shock effects on elF-4F. J . Biol. Chem. 262380-388. Durham, H.D., Pena, S . D . , and Carpenter, S. (1983): The neurotoxins 2.5-hexanedione and acrylamide promote aggregation of intermediate filaments in cultured fibroblasts. Muscle and Nerve 6:631-637. Eckert. B.S. ( 1985):Alteration of intermediate filament distribution in PtKl cells by acrylamide. Eur. J. Cell Biol. 37:169-174. Eckert, B.S. (1986): Alteration o f the distribution of intermediate filaments in PtKl cells by acrylamide 11: Effects on the organization o f cytoplasmic organelles. Cell Motil. Cytoskel. 6: 15-24. Eckert, B.S.. and Yeagle, P.L. (1987): NMR investigation of acrylainide-induced changes in intermediate filament organization. J . Cell Biol. 105:422a.
120
Aggeler and Seely
Eckert, B.S., and Yeagle, P.L. (1988): Acrylamide treatment of PtKl cells causes dephosphorylation of keratin polypeptides. Cell Motil. Cytoskel. 11:24-30. Franke, W.W., Schmid, E., Osborn, M., and Weber, K . (1978): Different intermediate-sized filaments distinguished by immunofluorescence microscopy. Proc. Natl. Acad. Sci. USA 75: 5034-5038. Green, K.J., Talian, J.C., and Goldman, R.D. (1986): Relationship between intermediate filaments and microfilaments in cultured fibroblasts: evidence for common foci during cell spreading. Cell Motil. Cytoskel. 6:406-418. Harris, E.D., Jr., Reynolds, J.J., and Werb, Z. (1975): Cytochalasin B increases collagenase production by cells in vitro. Nature 257:243-244. Hynes, R.O. (1987): Integrins: a family of cell surface receptors. Cell 48549-554, Laemmli, U.K. (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685. Lapadula, D.M., Irwin, R.D., Suwita, E., and Abou-Donia, M . B . (1986): Cross-linking of neurofilament proteins of rat spinal cord in vivo after administration of 2,5-hexanedione. J . Neurochem. 46:1843-1850. Laszlo, A,, and Bissell, M.J. (1983): TPA induces simultaneous alteration in the synthesis and organization of vimentin. Exp. Cell Res. 148:221-234. Ornelles, D.A., Fey, E.G., and Penman, S . (1986): Cytochalasin D releases mRNA from the cytoskeletal framework and inhibits protein synthesis. Mol. Cell Biol. 6: 1650-1662. Proud, C.G. (1986): Guanine nucleotides, protein phosphorylation and the control of translation. TIBS 11:73-77. Ruoslahti, E., and Pierschbacher, M.D. (1987): New perspectives in cell adhesion: RGD and integrins. Science 238:491-497. Sager, P.R. (1989): Cytoskeletal effects of acrylamide and 2,5,-hexanedione: selective aggregation of vimentin filaments. Toxicol. Appl. Pharmacol. 97:141-155. Schliwa, M. (1982): Action of cytochalasin D on cytoskeletal networks. J . Cell Biol. 92:79-91.
Schonthal, A . , Herrlich, P.. Rahmsdorf, H.J., and Ponta, H. (1988): Requirement for fos gene expression in the transcriptional activation of collagenase by other oncogenes and phorbol esters. Cell 54:325-334. Schorpp, M., Mallick, U . , Rahmsdorf, H.J., and Herrlich, P. (1984): UV-induced extracellular factor from human fibroblasts communicates the UV response to nonirradiated cells. Cell 37: 861-868. Steinert, P.M., Steven, A.C., and Roop, D.R. (1985): The molecular biology of intermediate filaments. Cell 42:41 1-419. Tilson, H.A. (1981): The neurotoxicity of acrylamide: An overview. Neurobehavior. Toxicol. Teratol. 3:445-461. Unemori, E.N., and Werb, Z. (1986): Reorganization of polymerized actin: a possible trigger for induction of procollagenase in fibroblasts cultured on and in collagen gels. J. Cell Biol. 103: 1021-1032. Unemori, E.N., and Werb, Z. (1988): Collagenase expression and endogenous activation in rabbit synovial fibroblasts stimulated by the calcium ionophore A23187. J . Biol. Chem. 263:1625216259. Vance, B.A., Kowalski, C.G., and Brinckerhoff, C.E. (1989): Heat shock of rabbit synovial fibroblasts increases expression of mRNAs for two metalloproteinases, collagenase and stromelysin. J. Cell Biol. 108:2037-2044. Werb, Z., and Aggeler, J. (1978): Proteases induce secretion of collagenase and plasminogen activator by fibroblasts. Proc. Natl. Acad. Sci. USA 75:1839-1843. Werb, Z., Hembry, R.M., Murphy, G., and Aggeler, J. (1986): Commitment to expression of the metalloendopeptidases, collagenasc and stromclysin: rclationship of inducing cvcnts to changes in cytoskeletal architecture. J. Cell Biol. 102:697702. Werb, Z., Tremble, P.M., Behrendtsen, O., Crowley, E., and Damsky, C.H. (1989): Signal transduction through the fibronectin receptor induces collagenase and stromelysin gene expression. J . Cell Biol. 109:877-890.
Cytoskeletal Dynamics in Rabbit Synovial Fibroblasts: 1. Effects of Acrylamide on Inte rmediate Filaments and Microfilaments Judith Aggeler and Keith Seely
Department of Human Anatomy, School of Medicine, University of California, Davis, California Rabbit synovial fibroblasts respond to changes in cell shape and cytoskeletal architecture by altering specific gene expression. We have tested the ability of acrylamide, a neurotoxin that alters the distribution of intermediate filaments in cultured PtKl cells, to induce metalloprotease expression in synovial fibroblasts. Cells treated with 2-20 mM acrylamide for 5 to 24 h underwent shape changes similar to cells treated with the tumor promoter phorbol myristate acetate. Intermediate filaments visualized with anti-vimentin antibodies did not collapse into a perinuclear cap in these rounded cells, but were still present in the extended cell processes. Unexpectedly, when actin was visualized in acrylamide-treated cells, extensive dissociation and clumping of microfilaments was observed. Concentrations of acrylamide > 10 mM were cytotoxic, but cells recovered completely after 24 h incubation with 5 mM acrylamide. Like other agents that alter cell shape and actin distribution in synovial fibroblasts, acrylamide also induced expression of the secreted metalloprotease collagenase. Although some recent evidence suggests that acrylamide may be able to exert its collagenase-inducing effects extracellularly , perhaps through transmembrane matrix receptors, our observation that this neurotoxin dramatically alters protein synthesis in synovial fibroblasts suggests that direct effects on cell metabolism may also play a role in acute acrylamide intoxication. Key words: vimentin, collagenase, cell shape, gene expression
INTRODUCTION The cytoskeleton of mammalian cells is made up of three major filament systems: microfilaments, intermediate filaments, and microtubules. A vast literature exists documenting the varied structure, function, and dynamics of actin filaments and microtubules. Although the underlying structure of intermediate filaments has recently been described in some detail [Steinert et al., 19851, the functions of this filament system have not yet been fully elucidated. This anomalous situation has arisen in part because the intermediate filament networks of most cells are relatively stable, making studies of filament assembly and turnover difficult, and in part because, unlike other cytoskeletal components, no agent has previously been available that can perturb intermediate filaments with reasonable specificity. Recent studm innn iaxi,..,
T
:-- T.,.
ies by Eckert [1985, 19861, have shown that the neurotoxin acrylamide can collapse the intermediate filament network of cultured PtK 1 cells without obvious effect on actin or microtubules. Such activities as vesicular movement and membrane ruffling continued in acrylamidetreated cells, indicating that intermediate filaments are not central to the mechanism(s) underlying these vital
Received June 22. 1989: accepted February 1. 1990. Address reprint requests to Dr. Judith Aggeler, Department o f Human Anatomy. University of California, Davis, CA 95616. Abbreviations used: DME. Dulhecco’s modification of Eagle‘s minimum essential medium: HBSS. Hanks’ balanced salt solution: PMA, phorhol myristate acetate: SDS-PAGE, sodium dodecylsulfate polyacrylamide gel electrophoresis.
Actin Dynamics in Acrylamide-Treated Cells
111
cell functions. These observations form the basis for further experimental investigation into the role of intermediate filaments in many cell functions that are thought to depend, at least in part. on components of the cellular cytoskeleton. We have been interested for a number of years in the mechanism(s) leading to the induction and sustained expression of metalloproteases (collagenase and stromelysin) in cultured fibroblasts [Werb and Aggeler, 1978; Aggeler et al., 1984a,b; Chin et al., 1985; Werb et al., 19861. A variety of agents with pharmacologic activities implicating plasma membrane receptors [Werb and Aggeler, 1978; Aggeler et al., 1984b; Werb et al., 19891, intracellular Ca mobilization [Aggeler et al., 1984b; Unemori and Werb, 19881, and protein kinases [Aggeler et al., 1984al have been identified as inducers. An important correlate to these studies has been the observation that induction of metalloprotease expression in cultured cells is preceded by changes in cell morphology (cell rounding) [Werb and Aggeler, 1978; Aggeler et al., 1984bl. In addition, it is clear that expression of collagenase and stromelysin is induced by a relatively short exposure to most pharmacological agents (1-6 h) and that once induced, synthesis and secretion can continue for many days in the absence of the agent [Aggeler et al., 1984bl. Present evidence strongly suggests that the cell shape changes associated with collagenase induction are based on alterations of the actin cytoskeleton [Aggeler et al., 1984b; Unemori and Werb, 19861 and that microtubules are not directly involved [Unemori and Werb, 19861. Intermediate filaments have not previously been implicated in this modulation of gene expression; however, the recent report of Green and colleagues 119861 that microfilaments and intermediate filaments interact during spreading of rounded cells in culture suggested to us that this filament system might also be involved in other cell functions previously thought to be solely actinbased. We now report our attempts to determine the role played by intermediate filaments in metalloprotease induction by treating cultured rabbit synovial fibroblasts with acrylamide. The accompanying paper [Aggeler, 19901 documents the redistribution of microfilaments and intermediate filaments during the period of sustained +
+
Fig. I . Acrylamide-induced shape changes in rabbit synovial fibroblasts. Rabbit synovial fibroblasts were incubated in the absence (a) or presence ofacrylamide (S m M ) for 6 (b), 8 ( c ) .or 16 h ( d ) .With time. acrylamide-treated cells retract from each other and become increasingly rounded. Open arrows indicate retraction fibers. These morphological changes art similar t o those observed in PMA-treated cells, where the degree o f shape change has been shown to predict the extent of induced collagenase synthesis (see Aggeler et al., 1984a. Fig. I ) . x 300.
112
Aggeler and Seely TABLE I. Acrylamide Inhibits Protein Synthesis* Protein synthesis Cells Treatment Control Acrylamide
Medium
Time (h)
Incorporation (cpm x lo4)
Inhibition
lncorporation (cpm x lo4)
Inhibition
(or,)
18
450 268 129
40.4 71.3
2.5 2.8 2.9
-
6 18
(%)
0 0
*Rabbit synovial fibroblasts incubated in serum-free medium in the absence or presence of 5 mM acrylamide for the times indicated were labeled for 2 h with ["SI-methionine. Total protein synthesis was determined by counting an aliquot of the detergent-solubilized cell layer. Biosynthesis of secreted protein was determined by precipitating from the medium with trichloroacetic acid and counting. Inhibition of protein synthesis by acrylamide is expressed as a percentage of the untreated control cell incorporation.
and maintained in Dulbecco's modified Eagle's medium (DME) supplemented with 10% fetal bovine serum (both C 6 2 4 4 8 from GIBCO, Grand Island, NY), as described previously [Aggeler et al., 1984al. Cells were subcultured weekly ( 1 5 ) with 0.25% trypsin and used between passages 2 and 10. Cells were plated onto coverslips (1 2 mm dia.) in 12-well plates (Costar) in DME plus serum 16 to 24 h before washing with Hanks' balanced salt solution (HBSS) and addition of serum-free DME supplemented with 0.2% lactalbumin hydrolysate. For most experiments cells were treated with acrylamide (electrophoresis grade; BioRad, Richmond, CA) in serum-free medium CL containing 0.2% lactalbumin hydrolysate for the times indicated. Serum-free medium has been used in this experimental system in the past to facilitate detection of collagenase activity. The morphological changes observed in synovial fibroblasts in the presence of acrylamide (see below) are also observed in cells treated in the presence of serum; however, recovery from treatment 1 2 3 4 after toxin removal is retarded in the absence of serum Fig. 2. Effects of acrylamide on synthesis of secreted proteins. Rabbit [Aggeler, 19901. Concentrations of acrylamide 5 5 mM synovial fibroblasts were incubated for 6 h in the absence (lane 1) or were not toxic to synovial fibroblasts, as judged by trypresence (lane 2-4) of 5 mM acrylamide, at which time they were pan blue exclusion (>95% of cells negative)(not shown) either radiolabeled immediately (lanes 1.2) or allowed to recover from acrylamide treatment in fresh serum-free medium for an additional 18 and by the rapid respreading of the cells after acrylamide (lane 3) or 42 h (lane 4)(total incubation 6, 24, or 48 h). This brief was removed [Aggeler, 19901. In come control experitreatment with acrylamide induces sustained procollagenase gene ex- ments, cells were treated with phorbol myristate acetate pression (CL) in respread cells after 24-48 h (lanes 3.4). The secreted (PMA)(20 nglml) or colchicine ( lop6 M)(Sigma Chempolypeptide migrating at M, = 53,000 has previously been identified ical Corp., St. Louis, MO), as described for acrylamide as procollagenase by immunoprecipitation and peptide mapping above. PMA was dissolved in ethanol at 1 mg/ml and [Aggeler, et al., 1984aj. stored at -20°C. Colchicine was dissolved in water at lo-' M and stored at 4°C. collagenase synthesis and secretion that follows withdrawal of inducing agents. Labeling of Proteins and Sodium Dodecylsulfate-PolyacrylamideGel MATERIALS AND METHODS Electrophoresis (SDS-PAGE) Cell Culture Cells plated into 12-well culture dishes (25 mm dia., Costar) and exposed to CB, CD, or acrylamide Rabbit synovial fibroblasts were harvested from the were pulse-labeled with 50 pCi/ml [%I-methionine knee joints of 4-6-week old New Zealand white rabbits x
-
Actin Dynamics in Acrylamide-Treated Cells
113
Fig. 3. Effects of acrylamide, PMA, and colchicine on intermediate filaments. Rabbit synovial fibroblasts were treated with acrylamide ( 5 mM) (a), PMA (20 ngiml) (b). or colchicine (lo-' M ) ( c ) for 16 h before fixation and staining for intermediate filaments (vimentin). Note that intermediate filament bundles appear to fill the numerous
retraction fibers present in both acrylamide- and PMA-treated cultures (open arrows). In contrast, colchicine (c) causes partial collapse of the intermediate filament network away from the cell periphery (arrowheads) into a perinuclear location. Unlike acrylamide and PMA, colchicine does not cause cell rounding o r retraction. X 300.
(1,350 Ci/mmol; Amersham Corp., Lisle, IL) in methionine-free DME. After 2 h, radiolabeled secreted protein was precipitated from the medium by the addition of cold trichloroacetic acid (final concentration 5 % ) . The washed protein pellets were resuspended in Laemmli sample buffer containing 0.5% P-mercapto-
ethanol and boiled for 3 min before electrophoresis [Laemmli, 19701. SDS-PAGE was carried out using 12% minigels with a 3% stacking gel. Dried gels were autoradiographed with Kodak X-Omat AR x-ray film, preflashed before exposure, as described previously [Aggeler et al., 1984al.
114
Aggeler and Seely
Fig. 4. Microfilament disassembly in acrylamide-treated cells. Rabbit synovial fibroblasts were incubated in serum-free medium in the absence (a)or presence ofacrylamide ( 5 mM) for S ( b ) or 18 h ( c ) before fixation and staining o f actin with rhodamine-phalloidin. At early times (b), a few treated cells have begun to retract. and some clumping of actin and decreased numbers of stress fibers are observed. even in
apparently spread cells (arrows). By 18 h (c) a majority of the cells have rounded to some extent and virtually all cells in the culture show some degree of actin clumping Some of the larger actin clumps are clearly visible in the accompanying phase contrast images of the same cells (arrowheads). X 300.
Actin Dynamics in Acrylamide-Treated Cells
115
Fig. 5 . Actin distribution in acrylamide-treated cells. Rabbit synovial fibroblasts were incubated in the absence (a) or presence (b) of acrylamide ( 5 mM) for 18 h before fixation and staining of actin with a monoclonal anti-actin antibody. Disassembly and clumping of microfilaments in acrylamide-treated cells is similar to that observed with rhodamine-phalloidin (Fig. 4).A moderate degree of diffuse cytoplas-
mic staining in both controls (a) and treated cells (b) indicates the presence of G-actin, which does not appear to be altered in acrylamide-treated cells. As in the rhodamine-phalloidin-stained cells (Fig. 4), the actin clumps are also visible in the phase contrast images of the same cells (arrowheads, b). X 300.
lmmunofluorescence Microscopy
sham). Intermediate filaments were visualized with a mouse monoclonal antibody raised against chicken gizzard vimentin (ICN Immunobiologicals, Lisle, IL), followed by biotinylated rabbit anti-mouse IgG (Amersham) and streptavidin-fluorescein (BRL, Bethesda, MD) or streptavidin-Texas Red. Cells were observed under epiilumination using a Zeiss Photomicroscope 111 equipped with a Plan NEOFLUAR 25 objective, N.A. 0.8, and photographed with Kodak Tri-X 400 film.
Fibroblasts cultured on 12 mm diameter glass coverslips were incubated with inducing agents for 1 to 24 h. Cells were then washed and fixed with freshly prepared 2% paraformaldehyde in phosphate buffered saline, pH 6.8, for 15 min at 22°C. Fixed cells were permeabilized by immersing them for 5 min in cold acetone (-20°C). Actin filaments were visualized with rhodamine-phalloidin (30 min, 22"C), as described by the manufacturer (Molecular Probes, Junction City, OR). In some experiments, actin was detected using a monoclonal antibody directed against chicken gizzard actin (Amersham Corp., Arlington Heights, IL), followed by biotinylated rabbit anti-mouse IgM and streptavidin-Texas Red (Amer-
Transmission Electron Microscopy
For transmission electron microscopy, rabbit synovial fibroblasts plated in 16 mm culture wells were incubated with 5 mM acrylamide in serum-free medium for
116
Aggeler and Seely
20 h. Cells were fixed with 2% glutaraldehyde (in Hanks' balanced salt solution containing 20 mM HEPES buffer, pH 6.8) for 1 hr at room temperature, postfixed with 0.5% OsO, for 15 min at 4"C, dehydrated, and embedded en bloc in EpodAraldite. Pieces of hardened Epon were then reembedded so that thin sections cut perpendicular to the dish surface could be prepared. Sections were observed using a Philips 400 electron microscope. RESULTS Alteration of Cell Shape in Rabbit Fibroblasts Treated With Acrylamide
Rabbit synovial fibroblasts treated with 2-20 mM acrylamide for up to 24 h showed dose-dependent changes in cell shape. This cell rounding was progressive with time, as illustrated in Figure 1 for cells treated with 5 mM acrylamide. After 5-6 h exposure, a few cells were rounded and showed spindly retraction fibers (Fig. lb). By 8 h, cell rounding and retraction were much more pronounced (Fig. lc), and after 16 h, virtually all cells in the culture showed morphological changes (Fig. Id). Essentially the same result was obtained for cells treated in the presence or absence of serum. Concentrations of acrylamide 95%) and by respreading and growth of cells after the drug was removed [Aggeler, 19901. Higher concentrations (10-20 mM) routinely killed cells, even after relatively brief exposures ( 5 h). This effective concentration range is similar to that observed previously for other cultured cells [Eckert, 19851. Acrylamide Induces Procollagenase Synthesis and Secretion
We have previously shown that a variety of agents that cause cell shape changes in rabbit synovial fibro-
Fig. 6. Thin section views of acrylamide-induced actin clumps. Rabbit synovial fibroblasts were incubated with 5 mM acrylamide for 20 h, fixed, and embedded prior to sectioning. A low magnification view of an entire cell (a) reveals that large aggregates of actin are located in the perinuclear cytoplasm (arrow), well above the basal surface o f the cell. At high magnification (b), this aggregate is seen to consist of numerous small dense clumps of thin filaments (open arrows). Cell organelles appear to be excluded from these aggregates, and even intermediate filaments. numerous in the surrounding cytoplasm (closed arrow). are not observed within these areas. In (c). an actin aggregate at the basal surface (arrowheads) of a cell treated with cytochalasin D for 20 h is shown for comparison. m = mitochondrion. a: Bar = 2 p m , X6.900. b.c: Bar = 0.2 pm, X90.000.
Actin Dynamics in Acrylamide-Treated Cells
blasts cause these cells to undergo a specific switch in gene expression characterized by down regulation of collagen and fibronectin synthesis and induction of the extracellular matrix-degrading metalloproteases collagenase and stromelysin [Werb and Aggeler, 1978; Aggeler et al., 1984a,b; Unemori and Werb, 19861. When newly synthesized secreted proteins of acrylamide-treated fibroblasts were examined by SDS-PAGE, a similar induction of collagenase expression was observed (Fig. 2). Because protein synthesis was strongly inhibited (40-70% reduction in ["S]-methionine incorporation into cellular protein) in the presence of acrylamide (Table I), induction of collagenase was most clearly shown in cells that had respread after removal of the drug.
Intermediate Filament Distribution in Acrylamide-Treated Cells
Because cultured PtKl cells exposed to S mM acrylamide show dramatic alterations in the distribution of the intermediate filament cytoskeleton (both vimentin and cytokeratin)[Eckert, 19851, and because both cell shape changes and collagenase induction were observed in acrylamide-treated synovial fibroblasts, we were particularly interested in determining whether the intermediate filaments of these cells were also affected by the drug. When the intermediate filament mesh of acrylamide-treated rabbit synovial fibroblasts was observed by immunofluorescence microscopy using an anti-vimentin antibody, wholesale collapse and retraction of intermediate filaments was not observed (Fig. 3a), although some aggregation of individual filaments into bundles may occur. These bright-staining bundles of vimentin filaments were present throughout the cytoplasm of rounded cells and extended out into the numerous retraction fibers. This pattern of intermediate filament distribution was qualitatively similar to that observed in cells treated with another collagenase-inducing agent, the phorbol ester tumor promoter, PMA (Fig. 3b), as described previously by Laszlo and Bissell [1983]. In contrast, the microtubule-depolymerizing alkaloid, colchicine, neither alters the overall morphology or microfilament distribution of rabbit synovial fibroblasts [Werb and Aggeler, 1978; Unemori and Werb, 19861 nor induces collagenase [Unemori and Werb, 19861; however, it does cause collapse of intermediate filaments toward the nucleus (Fig. 3c), as has been described previously for a variety of cultured cells [Franke et al., 19781. These results indicated that the dramatic changes in cell shape observed in acrylamide-treated fibroblasts were not correlated with distinctive changes in intermediate filament distribution.
117
TABLE 11. Acrylamide-Induced Changes in Actin Cytoskeleton* Morphology Treatment ~~
Time (h) Unchanged
Control Acrylamide (mM) 2
5
24
97
3
0
5 18 24
95 33 12
4 48 36
19 52
5
63 0 0
32 0 I1
4 100 89
18
24 ~~
Intermediate ClumDed
~
1
~
"Rabbit synovial fibroblasts treated with acrylamide for the times and at the concentrations indicated were fixed and stained with rhodamine phalloidin to visualize actin filaments. One hundred randomly-chosen cells were scored at the microscope for each treatment; values represent numbers of cells scored in each category. Moderate changes in cell morphology and actin cytoskeleton were characterized by some cytoplasmic retraction and condensation of filamentous actin (for example, see Fig. 4b). Rounded cells showed no sign of stress fibers and were characterized by localization of actin in several large cytoplasmic clumps.
Alteration of Actin Cytoskeletal Architecture in Rabbit Fibroblasts Treated With Acrylamide
Having found no major alterations of the intermediate filament mesh of acrylamide-treated fibroblasts, we next sought to determine the effects of acrylamide on the actin cytoskeleton of these cells. In contrast to intermediate filaments, the pattern of actin filament distribution was dramatically altered in cells rounded by exposure to acrylamide (Fig. 4). In parallel to the changes in cell shape observed in Figure 1, acrylamide-treated cells showed progressive condensation and clumping of filamentous actin, accompanied by loss of distinctive microfilament bundles or stress fibers. With time, increasing numbers of cells in these cultures showed these changes in actin distribution (Table II), and by 16-18 h of treatment with S mM acrylamide, virtually every cell displayed some degree of actin redistribution. In most cells, actin was completely collapsed into a number of bright-staining aggregates (Fig. 4c). When acrylamidetreated cells were stained using an anti-actin antibody, the same pattern of actin clumping was observed (Fig. 5b). Cells also displayed a moderate level of diffuse cytoplasmic actin-staining (G-actin), which appeared to be somewhat reduced by treatment with acrylamide (Fig. Sb). Comparison of the phase and fluorescence images of the same cells shows that the actin clumps are clearly visible by phase optics (arrowheads, Figs. 4c and 5b). Interestingly, staining of these clumps with the anti-actin antibody often appeared restricted to their surface (arrows, Fig. 5b), as if the antibody did not penetrate the aggregate, whereas rhodamine-phalloidin staining usually appeared uniform throughout the clumps.
118
Aggeler and Seely
Unlike the basally located microfilament foci induced by cytochalasin treatment (Fig. 6c) [Schliwa, 19821, the actin clumps induced by acrylamide appeared to be located in the apical cytoplasm of rounded cells, well above the plane of focus of the culture dish. This impression was confirmed by transmission electron microscopy (Fig. 6a). High magnification views indicated that these foci consisted of loose aggregates of small, dense actin clumps (Fig. 6b). Numerous intermediate filaments were found in the cytoplasm surrounding these aggregates but did not appear to be present within the clumps themselves. DISCUSSION
Although the structure and distribution of intermediate filaments have been described in some detail, their function(s) in cells have remained unclear. We have attempted to use the neurotoxin acrylamide, which alters the intermediate filament distribution of certain cultured cell types, to investigate the role played by these filaments in rabbit synovial fibroblasts. Unlike the response described for PtKl [Eckert, 1985, 19861 and PtK2 cells [Sager, 19891, the vimentin meshwork of synovial fibroblasts was not specifically affected by acrylamide. In addition, except under conditions where toxic concentrations of drug were employed ( > l o mM), we have not observed collapse of the cytokeratin filament mesh of cultured primary mammary epithelial cells treated with acrylamide either (Seely and Aggeler, in preparation). At present, we do not know why the intermediate filament networks of fibroblasts and epithelial cells in our culture systems display this stability. Both of these models use primary cells, rather than cell lines, and it may be that the transformation event(s) necessary to produce immortal lines such as PtKl alter the cytoskeleton in subtle ways. Alternatively, Eckert and Yeagle [Eckert and Yeagle, 19881 have recently shown that acrylamide causes dephosphorylation of keratins in PtKl cells, and that intermediate filament collapse can be prevented in these cells by raising intracellular levels of cyclic AMP, thus, presumably, stimulating endogenous CAMP-dependent kinase activity. Experiments are in progress to determine whether intermediate filaments in our primary cells are protected from collapse by naturally high levels of intracellular CAMP under our culture conditions. One reason for our particular interest in the intermediate filament network of synovial fibroblasts was the possibility that microfilaments and intermediate filaments might be interacting to regulate metalloprotease gene expression in these cells. Green and colleagues [Green et al., 19861 have reported that microfilaments and intermediate filaments reform from common foci in cultured cells respreading after rounding by trypsin and
cytochalasin, two agents that also induce collagenase gene expression in cultured synovial fibroblasts [Werb and Aggeler, 1978; Aggeler et al., 1984b; Harris et al., 19751. Nevertheless, as shown in the accompanying report [Aggeler, 19901, no indication of such cytoskeletal interaction was observed in synovial fibroblasts respreading after treatment with trypsin, cytochalasins B or D, or acrylamide. Taken together, our present results provide little indication for a distinct role for intermediate filaments in metalloprotease icduction and expression in cultured rabbit synovial fibroblasts and serve to reinforce the importance of the actin cytoskeleton in these processes. In contrast to the apparent lack of effect on intermediate filaments, acrylamide caused a dramatic alteration of the actin cytoskeleton of cultured rabbit synovial fibroblasts. Similar, but less dramatic, alterations in stress fibers have been noted in cultured human skin fibroblasts treated with 2 mM acrylamide for 7 days, but not in cells treated with another neurotoxin, 2,5-hexanedione [Durham et al., 19831. More recently, Sager [1989] has shown both reduction of stress fibers and actin clumping in PtK2 cells treated with acrylamide or 2,5-hexanedione, but these changes were not accompanied by cell rounding. In the present study, marked cell rounding was accompanied by disappearance of stress fibers and collapse of microfilaments into perinuclear clumps. These loose aggregates of actin were unlike those caused by cytochalasin treatment, which are typically found subadjacent to the basal plasma membrane [Schliwa, 19821; nor did they resemble the dense clumps observed by Green et al. [1986] in trypsin-treated chick embryo fibroblasts, in that intermediate filaments were not intimately associated with them. The molecular basis for the collapse of the actin cytoskeleton in acrylamidetreated synovial fibroblasts is not known, but one intriguing possibility is that this drug is acting on plasma membrane receptors for extracellular adhesion proteins, such as fibronectin or vitronectin. Eckert and Yeagle 119871 have presented preliminary data indicating that acrylamide may not enter cells. Acrylamide is known to be a very reactive molecule [Tilson, 198 11, capable of interacting with sulfhydryl groups, such as those present in the extracellular domains of the integrins [Hynes, 1987; Ruoslahti and Pierschbacher, 19871, and of cross-linking proteins [Lapadula et al., 19861. Perhaps acrlamide can bind to and alter the conformation of cell surface receptors in such a way that associated cytoplasmic anchors for actin are disrupted [Damsky et al., 1979; Burridge et al., 19881. Such a mode of action would be consistent with our observation that acrylamide-treated cells respread more rapidly in the presence of a source of exogenous adhesion molecules, such as serum [Aggeler, 19901. In addition, recent evidence indicates that colla-
Actin Dynamics in Acrylamide-Treated Cells
genase can be induced in synovial fibroblasts by antibodies directed against cell-surface fibronectin receptors [Werb et al., 19891, another indication that acrylamide may be interacting with this transmembrane signal transduction system. The effects of acrylamide on overall protein synthetic rates in synovial fibroblasts are less easy to reconcile with an extracellular site of action, although inhibition of synthesis may simply be a consequence of the profound alteration of the actin cytoskeleton, as has been observed previously in some cytochalasin-treated cells [Ornelles et al., 1986; Blum and Wicha, 19881 (Seely and Aggeler, in preparation). In these cases, it has been suggested that translation is impeded because active polyribosomes are detached from the cytoskeleton and destabilized. Alternatively, inhibition of protein synthesis in acrylamide-treated fibroblasts may be due to alterations in the state of phosphorylation of components of the synthetic machinery. Several recent studies [DeBenedetti and Baglioni, 1986; Duncan et al., 1987; Duncan and Hershey, 19891 have suggested that phosphorylation of initiation factor(s) by a cGMP-dependent kinase activity [Proud, 19861 may contribute to the inhibition of normal protein synthesis observed in heat-shocked cells. Interestingly, collagenase itself is induced by heat shock [Schorpp et al., 1984; Vance et al., 19891. Thus, acute acrylamide intoxication may invoke some aspects of a more generalized stress response in rabbit synovial fibroblasts [Angel et al., 1986; Schonthal et al., 19881. Whatever the basis for the acute inhibition of fibroblast protein synthesis by acrylamide, it is rapidly reversible. Moreover, although overall protein synthesis can be dramatically reduced, synthesis of secreted proteins continues unchanged and induction of collagenase gene expression proceeds. Thus although acrylamide has not provided the hoped-for tool with which to identify a role for intermediate filaments in metalloprotease induction, it has presented us with an additional means for studying both induction of specific gene expression and mechanisms controlling synthesis of secreted proteins in these cells. In addition, the varied responses of microfilaments and intermediate filaments of different cell types to these neurotoxins opens fertile ground for testing the possible role of factors such as extracellular matrix and phosphorylation in mediating cytoskeletal function. ACKNOWLEDGMENTS
The authors would like to thank Dr. Fern Tablin, University of California, Davis, California, for many helpful discussions and for critical reading of this manuscript. Electron micrographs were produced in the Electron Microscopy Laboratory of the School of Medicine, University of California, Davis, California. This work
119
was supported in part by a grant from the University of California Cancer Research Coordinating Committee and by NIH Biomedical Research Support Grant S07-RR05684. Some of these results have appeared previously in abstract form [Aggeler and Seely, 19871. REFERENCES Aggeler, J . ( 1990): Cytoskeletal dynamics in rabbit synovial fibroblasts. 11. Reformation of stress fibers in cells rounded by treatment with collagenase-inducing agents. Cell Motil. Cytoskel. 16: I2 1- 132. Aggeler. J . , Frisch, S . M . . and Werb, Z. (1984a): Collagenase is a major gene product of' induced rabbit synovial fibroblasts. J . Cell Biol. 98:1656-1661. Aggeler. J.. Frisch. S.M., and Werb, Z. (1984b): Changes in cell shape correlate with collagenase gene expression in rabbit synovial fibroblasts. J . Cell Biol. 98:1662-1671. Aggeler. J., and Seely, K. (1987): Effects of acrylamide on the cytoskeletons of cultured fibroblasts and epithelial cells. J . Cell Biol. 105:208a. Angel. P., Poting, A. Mallick. U., Rahmsdorf, H.J., Schorpp, M., and Herrlich, P. (1986): Induction of metallothionein and other mRNA species by carcinogens and tumor promoters in primary human skin fibroblasts. Mol. Cell Biol. 6: 1760-1766. Blum, J.L., and Wicha, M.S. (1988): Role of the cytoskeleton in laminin induced mammary gene expression. J . Cell Physiol. 135: 13-22. Burridge, K., Fath, K., Kelly, T., Nucholls, G . , and Turner, C . ( 1988): Focal adhesions: transmembrane junctions between the extracellular matrix and the eytoskeleton. Ann. Rev. Cell Biol. 4:487-525. Chin, J.R., Murphy, G., and Werb, Z. (1985): Stromelysin. A connective tissue-degrading metalloendopeptidase secreted by stimulated rabbit synovial fibroblasts in parallel with collagenase. J. Biol. Chem. 260:12367-12376. Damsky, C . H . , Wylie, D.E., and Buck, C.A. (1979): Studies on the function of cell surface glycoproteins. 11. Possible role of surface glycoproteins in the control of cytoskeletal organization and surface morphology. J . Cell Biol. 80:403-415. DeBenedetti, A , , and Baglioni, C . (1986): Activation of hemin-regulated initiation factor-2 kinase in heat-shocked HeLa cells. J. Cell Biol. 261:338-342. Duncan, R.F., and Hershey. J.W.B. (1989): Protein synthesis and protein phosphorylation during heat stress, recovery, and adaptation. J. Cell Biol. 109:1467-1481. Duncan, R . , Milburn, S . C . , and Hershey, J.W.B. (1987): Regulated phosphorylation and low abundance of HeLa initiation factor elF-4F suggest a role in translational control: heat shock effects on elF-4F. J . Biol. Chem. 262380-388. Durham, H.D., Pena, S . D . , and Carpenter, S. (1983): The neurotoxins 2.5-hexanedione and acrylamide promote aggregation of intermediate filaments in cultured fibroblasts. Muscle and Nerve 6:631-637. Eckert. B.S. ( 1985):Alteration of intermediate filament distribution in PtKl cells by acrylamide. Eur. J. Cell Biol. 37:169-174. Eckert, B.S. (1986): Alteration o f the distribution of intermediate filaments in PtKl cells by acrylamide 11: Effects on the organization o f cytoplasmic organelles. Cell Motil. Cytoskel. 6: 15-24. Eckert, B.S.. and Yeagle, P.L. (1987): NMR investigation of acrylainide-induced changes in intermediate filament organization. J . Cell Biol. 105:422a.
120
Aggeler and Seely
Eckert, B.S., and Yeagle, P.L. (1988): Acrylamide treatment of PtKl cells causes dephosphorylation of keratin polypeptides. Cell Motil. Cytoskel. 11:24-30. Franke, W.W., Schmid, E., Osborn, M., and Weber, K . (1978): Different intermediate-sized filaments distinguished by immunofluorescence microscopy. Proc. Natl. Acad. Sci. USA 75: 5034-5038. Green, K.J., Talian, J.C., and Goldman, R.D. (1986): Relationship between intermediate filaments and microfilaments in cultured fibroblasts: evidence for common foci during cell spreading. Cell Motil. Cytoskel. 6:406-418. Harris, E.D., Jr., Reynolds, J.J., and Werb, Z. (1975): Cytochalasin B increases collagenase production by cells in vitro. Nature 257:243-244. Hynes, R.O. (1987): Integrins: a family of cell surface receptors. Cell 48549-554, Laemmli, U.K. (1970): Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685. Lapadula, D.M., Irwin, R.D., Suwita, E., and Abou-Donia, M . B . (1986): Cross-linking of neurofilament proteins of rat spinal cord in vivo after administration of 2,5-hexanedione. J . Neurochem. 46:1843-1850. Laszlo, A,, and Bissell, M.J. (1983): TPA induces simultaneous alteration in the synthesis and organization of vimentin. Exp. Cell Res. 148:221-234. Ornelles, D.A., Fey, E.G., and Penman, S . (1986): Cytochalasin D releases mRNA from the cytoskeletal framework and inhibits protein synthesis. Mol. Cell Biol. 6: 1650-1662. Proud, C.G. (1986): Guanine nucleotides, protein phosphorylation and the control of translation. TIBS 11:73-77. Ruoslahti, E., and Pierschbacher, M.D. (1987): New perspectives in cell adhesion: RGD and integrins. Science 238:491-497. Sager, P.R. (1989): Cytoskeletal effects of acrylamide and 2,5,-hexanedione: selective aggregation of vimentin filaments. Toxicol. Appl. Pharmacol. 97:141-155. Schliwa, M. (1982): Action of cytochalasin D on cytoskeletal networks. J . Cell Biol. 92:79-91.
Schonthal, A . , Herrlich, P.. Rahmsdorf, H.J., and Ponta, H. (1988): Requirement for fos gene expression in the transcriptional activation of collagenase by other oncogenes and phorbol esters. Cell 54:325-334. Schorpp, M., Mallick, U . , Rahmsdorf, H.J., and Herrlich, P. (1984): UV-induced extracellular factor from human fibroblasts communicates the UV response to nonirradiated cells. Cell 37: 861-868. Steinert, P.M., Steven, A.C., and Roop, D.R. (1985): The molecular biology of intermediate filaments. Cell 42:41 1-419. Tilson, H.A. (1981): The neurotoxicity of acrylamide: An overview. Neurobehavior. Toxicol. Teratol. 3:445-461. Unemori, E.N., and Werb, Z. (1986): Reorganization of polymerized actin: a possible trigger for induction of procollagenase in fibroblasts cultured on and in collagen gels. J. Cell Biol. 103: 1021-1032. Unemori, E.N., and Werb, Z. (1988): Collagenase expression and endogenous activation in rabbit synovial fibroblasts stimulated by the calcium ionophore A23187. J . Biol. Chem. 263:1625216259. Vance, B.A., Kowalski, C.G., and Brinckerhoff, C.E. (1989): Heat shock of rabbit synovial fibroblasts increases expression of mRNAs for two metalloproteinases, collagenase and stromelysin. J. Cell Biol. 108:2037-2044. Werb, Z., and Aggeler, J. (1978): Proteases induce secretion of collagenase and plasminogen activator by fibroblasts. Proc. Natl. Acad. Sci. USA 75:1839-1843. Werb, Z., Hembry, R.M., Murphy, G., and Aggeler, J. (1986): Commitment to expression of the metalloendopeptidases, collagenasc and stromclysin: rclationship of inducing cvcnts to changes in cytoskeletal architecture. J. Cell Biol. 102:697702. Werb, Z., Tremble, P.M., Behrendtsen, O., Crowley, E., and Damsky, C.H. (1989): Signal transduction through the fibronectin receptor induces collagenase and stromelysin gene expression. J . Cell Biol. 109:877-890.
