Antibody staining of 10-nm (100-Å) filaments in cultured smooth, cardiac and skeletal muscle cells

Round smooth-surfaced filaments approximately 10 nm in diameter (range 8–12 nm) are now recognized as a ubiquitous and distinct class of cytoplasmic filaments. These filaments, commonly referred to as 10-nm or 100-Å filaments, have been observed in a large number of different cell types, including fibroblasts (Ishikawa, Bischoff & Holtzer, 1969; Goldman & Knipe, 1973), leukaemia cells (Felix & Strâuli, 1976) HeLa cells (Lenk, Ransom, Kaufmann & Penman, 1977), epithelial cells (Ishikawa et al. 1969), endothelial cells (Uehara, Campbell & Burnstock, 1971 ; Yohro & Bumstock, 1973), skeletal muscle (Ishikawa, Bischoff & Holtzer, 1968; Kelly, 1969; Uehara, Campbell & Burnstock, 1976), cardiac muscle (Rash, Biesele & Gey, 1970; Ferrans & Roberts, 1973 a), Purkinje fibres (Thornell, 1974; Oliphant & Loewen, 1976), and nerve (Shelanski, Albert, DeVries & Norton, 1971). Their morphological identity with tonofilaments of epithelial cells (Kelly, 1966) early suggested their involvement in structural support, although they have also been implicated in processes such as intracellular transport of organelles (Goldman & Follett, 1969) and in axoplasmic transport in neurons (Huneeus & Davison, 1970).

The addition of mitotic inhibitors such as colchicine to a wide variety of cells causes a juxtanuclear accumulation of 10-nm filaments (Ishikawa et al. 1968; Wisniewski, Shelanski & Terry, 1968; Blose & Chacko, 1976; Starger & Goldman, 1977) which are seen either as contorted biréfringent bands or a compact ‘cap’. The significance of this accumulation of 10-nm filaments under conditions in which mitogenic factors are involved is unknown. This type of modification is particularly interesting however in light of the increase in numbers of 10-nm filaments in some cells in pathological conditions such as cardiac myxomas (Ferrans & Roberts, 1973 b), neuroblastoma (Bertolini et al. 1977), human breast tumours (Tumilowicz & Sarkav, 1972), hypertrophied smooth muscle caused by stenosis of the ileum (Gabella, 1975), rhabdomyoblastomas (Bôcker & Stegner, 1975), and leiomyosarcomas (Bôcker & Strecker, 1975).

Antibody staining techniques have been used successfully in conjunction with fluorescence and electron microscopy to gain additional information about the location of proteins in the skeletal muscle myofibril (see Pepe, 1975; Offer, 1976), and have been used extensively in studies of the organization of contractile proteins in various types of cultured cells (e.g. see Lazarides, 1975 a, b, 1976; Goldman, Lazarides, Pollack & Weber, 1975). We have been applying this method to the study of smooth muscle cells in culture in an attempt to gain further information on the organization of contractile and cytoskeletal elements in smooth muscle in vivo under normal and pathological conditions. The localization of the contractile proteins actin, myosin and tropomyosin (Grôschel-Stewart, Chamley, Campbell & Bumstock, 1975; Chamley, Grôschel-Stewart, Campbell & Bumstock, 1977; Chamley, Campbell, McConnell & Grôschel-Stewart, 1977; Chamley-Campbell, Campbell, Grôschel-Stewart & Bumstock, 1977) has already been examined. With the recent finding that a polypeptide of around 55000 mol. wt constituted the major component of the prominent 10-nm filaments in smooth muscle (Cooke, 1976; Small & Sobieszek, 1977a) it became feasible to complement these studies with investigations of the distribution of the 10-nm filaments. Accordingly, we present here the results obtained with smooth muscle and with other muscle and non-muscle cells using an antibody raised against the 55000 mol. wt polypeptide of the 10-nm filaments.

During the course of these studies other reports appeared in which antibodies were either raised against the filament polypeptide (Lazarides & Hubbard, 1976), or native filaments (Blose, Shelanski & Chacko, 1977; Starcer & Goldman, 1977). Autoantibodies were also found in human disorders (Kurki, Linder, Virtanen & Stenman, 1977) or in normal animals (Osborn, Franke & Weber, 1977; Gordon, Bushnell & Burridge, 1978 ; Hynes & Destree, 1978). None of these reports, however, dealt with the filament organization in smooth muscle cells. In addition, it was apparent from these various data that some degree of species and tissue specificity existed, pointing to differences between 10-nm filaments in different cell types. Our own results also point to differences in the specificity of the antibody raised against the 55000 Dalton component of gizzard; these differences are discussed in relation to ultrastructural studies of muscle tissue and to the possible number of classes of 10-nm filaments.

Cardiac muscle from newborn rats and 10-day-old chick embryos, skeletal muscle from newborn rats and 19-day chick embryos, smooth muscle from 10-day chick embryo gizzards, newborn guinea-pig vas deferens and taenia coli, 5-month-old rabbit aorta and 6o-year-old human saphenous vein, and endothelium from human umbilical vein were dispersed into single cells by the use of enzymes (see Mark, Chamley & Bumstock, 1973 ; Campbell, Chamley & Burnstock, 1974; Chamley et al. 1977). Cells were resuspended in medium 199 to a concentration of 10⁵ cells per ml and grown on glass coverslips in modified Rose chambers. Expiants of newborn guinea-pig sympathetic ganglia and strips of Auerbach’s plexus from taenia coli were grown on poly-lysine-coated coverslips (see Chamley, Mark, Campbell & Burnstock, 1972).

In some cultures, colchicine (Sigma) at a concentration of 10⁻⁶ M was added to the medium at both 2 days and 4 h prior to fixation of the cells.

Isolated smooth muscle cells from adult guinea-pig taenia coli and vas deferens, and smooth muscle cell ‘ghosts’ from adult guinea-pig taenia coli were prepared as described elsewhere (Small & Sobieszek, 1977a, b;Small, 1977).

Tissue was frozen in liquid nitrogen-isopentane and 4-μm sections cut on a Slee cryostat.

Cultures of smooth muscle cells were fixed in phosphate-buffered 5% glutaraldehyde for i h, washed overnight in phosphate buffer, then further fixed for 1 h in phosphate-buffered 1% OsO₄. The material was then stained in a saturated solution of uranyl acetate in water, embedded in Araldite and sectioned as described elsewhere (Campbell et al. 1974). Sections were viewed with a Siemens 101 electron microscope.

A crude preparation of the 10-nm filament protein was prepared essentially following the method of Cooke (1976). The final precipitate was submitted to gel electrophoresis (10% polyacrylamide, 01% SDS) according to Weber & Osborn (1969) and the molecular weights of the protein bands calculated using calibration proteins from Boehringer Mannheim (‘Combithek’). The band corresponding to 55000 subunit molecular weight was cut out from the Coomassie Blue-stained gels and treated as follows. The gel-disks were ground in a mortar with quartz-sand in the presence of 0.05 M phosphate buffer — 0.1% SDS, pH 7.2. After sonication for 2 min in an ultra-sound water bath, the mixture was kept at 37 °C for 16–24 h (modified after Bray, 1973). After centrifugation, the soluble material was dialysed for 2 days against several changes of phosphate-buffered saline (PBS), and stored in aliquots (approximately 80-100 /4g of protein each) at —20 °C.

Rabbits were immunized with 3 × 100 μg of the eluate per rabbit at 21-day intervals, the antigen being injected intradermally and subcutaneously at multiple sites after emulsification with Freund’s complete adjuvant and Bordetella pertissus antigen. The presence of precipitating antibodies was then shown by immunodiffusion test, using 1% agar gels in barbiturate buffer, pH 8 2, containing 01% SDS (Obinata, Hasegawa, Masaki & Hayashi, 1976). In Fig. 1, the crude urea extract from the muscle homogenate (holes) was allowed to diffuse against the antiserum wells. In addition, an unstained and unfixed SDS-polyacrylamide gel was sliced lengthwise and placed on the centre, of the agar gel, 4–5 mm away from the antiserum well. A single precipitation line was found after 24 h incubation in a moist chamber (Fig. 2). After the position of the SDS gel on the agar had been marked with a diamond pencil, the agar gel was rinsed and dried. Both gels were then fixed and stained with Coomassie Brilliant Blue. When the SDS gel was replaced in the original position, the immunoprecipitation band corresponded to the 55000 molecular weight protein (Fig. 3).

The specificity of the 10-nm filament antiserum was determined by the enzyme-linked immunosorbent assay (ELISA) (Voller, Bidwell & Bartlett, 1976). Polystyrene tubes were sensitized by adding 0.3 ml of antigen in carbonate buffer (pH 9 6) and incubating the tubes at 4 °C overnight. The tubes were washed 3 times with phosphate-buffered saline containing 2% Tween 20. 0 3 ml of rabbit antiserum to 10-nm filament protein was added to the tubes and left at room temperature for 5 h. The tubes were washed again and 0.3 ml of anti-rabbit immunoglobulin conjugated with alkaline phosphatase (Sigma, P-4502) was added to the tubes and left at 4 °C overnight. After washing, 0.3 ml of enzyme substrate (4-nitrophenyl phosphate, 1 mg/ml) in 10% diethanolamine buffer (pH 9.8) was added to the tubes and left for 1h. The reaction was stopped by adding 0.5 ml 0.5 M NaOH, and the absorbance of the contents of the tubes were read in a spectrophotometer at 399 nm. The antigens employed are shown in Table 1.

Purified actomyosin from pig stomach was prepared according to the method of Small & Sobieszek (1977b). To demonstrate that the 53000 Dalton protein antibody was not contaminated with antibodies against actomyosin, 20 mg of the antibody dissolved in 3 ml PBS were adsorbed onto 60 mg of purified pig actomyosin at room temperature for i h, with frequent shaking. The suspension was then centrifuged at 500 g and the supernatant containing the antibody removed and used for staining.

The protein was prepared and purified according to the method of Small & Sobieszek (1977a). Prior to use it was dialysed against distilled water for 2 days and the pellet washed in PBS. 20 mg of the antibody were then absorbed on an excess of the purified protein for 1 h at room temperature, centrifuged, and the supernatant used for staining.

After varying times in culture, the chambers were disassembled and the coverslip on which the cells were grown rinsed in 2 changes of PBS and fixed in 2% formaldehyde for 20 min at room temperature. The coverslip was then washed in 2 changes of PBS for a total of 15 min and placed in acetone at − 15 °C for 10 min before air-drying and storage in a desiccator at 4 °C.

A small quantity of untreated antibody, or actomyosin, or 55000 Dalton protein-absorbed antibody, was diluted in PBS to a concentration of 2 mg/ml and applied to the fixed cells for 30 min in a moist Petri dish at room temperature. The coverslip was then washed with 2 changes of PBS for a total of 30 min and incubated for 30 min with a 1:15 dilution of fluorescein-conjugated IgG goat anti-rabbit gamma globulin (Cappel Lab.). Both the first and second antibodies were absorbed with chicken liver acetone powder (Sigma) prior to use. The coverslip was then washed in 2 changes of PBS for a total of 1 h and mounted in glycerol-0.1 M glycine buffer, pH 8.6 (7:3). A Zeiss Photomicroscope 111 fitted with epifluorescence optics and interference blue filter combination 455 and 490 was used for examining and photographing the cultures.

With each fixation, staining procedure, and cell type employed, control staining was carried out using pre-immune serum at a protein concentration 2 mg/ml.

Isolated cells and cell ‘ghosts’. A drop of PBS containing a suspension of isolated smooth muscle cells or cell ghosts was smeared on a microscope slide, rinsed briefly in distilled water and stained as for cultures (see Small & Sobieszek, 1977a).

Smooth muscle cells were grown in culture dishes containing coverslips (8 mm × 4 mm) carrying sterile silver electron-microscope grids coated with a plastic-carbon support film (Small & Celis, 1978 a). After appropriate times in culture the coverslips were removed, rinsed in PBS, treated with 0.1% Triton X-100 and extracted of actomyosin as described elsewhere (Small & Celis, 1978b). They were examined in the electron microscope, after negative staining.

Enzyme-dissociated smooth muscle cells within the first few days in culture are ribbon- or spindle-shaped (100–200 μm × 10–15 μm), with an oval nucleus containing 2 or more small, pale nucleoli. Their cytoplasm is phase dense and relatively free of visible inclusions and many contract spontaneously at rates of 1–7 per min. Cultures of vas deferens, taenia coli and chicken gizzard always contain fibroblasts in addition to smooth muscle cells, while adult aorta and saphenous vein can be grown as pure smooth muscle or with endothelial cells and /or fibroblasts (see Chamley et al. 1977).

In the first few days in culture after flattening, smooth muscle cells from adult human saphenous vein, adult rabbit aorta and newborn guinea pig vas deferens and taenia coli resemble in the electron microscope the in vivo cells (Fig. 4) and contain large bundles of myofilaments with dark bodies throughout. Associated with, and apparently passing between these dark bodies, are ‘10-nm filaments’. The filaments often appear randomly organized with respect to the myofilament bundles (see Uehara et al. 1971). In immature smooth muscle cells, such as those from the 10-day embryo chicken gizzard (Campbell, Uehara, Mark & Bumstock, 1971), the number of 10-nm filaments appears to be greater and they are often in the form of large randomly organized networks, although an association with electron-dense bodies can still be observed (Fig. 5).

With the antibody against the 55000 Dalton protein, all smooth muscle cells, regardless of their source and phenotypic state, stained brightly. The staining was intense in the cytoplasmic region around the nucleus and was reduced at the periphery (Fig. 6). Examination of the peripheral flattened areas at higher magnification revealed a number of fine fibrils which, although passing through the long axis of the cell, appeared to be randomly organized with respect to it (Fig. 7). Exactly the same pattern of staining was observed using the antibody after adsorption on purified actomyosin. In contrast, there was no staining after absorption of the antibody on the purified 55000 Dalton protein. Experiments with the pre-immune serum showed it to be negative when used as the first antibody.

Smooth muscle cells grown on electron-microscope grids, treated with Triton X-100 and extracted of actomyosin and tubulin with buffers of high and low salt concentration stained in the same way (Fig. 9) as non-extracted cells. At higher magnification and after negative staining these cells exhibited predominantly smooth-surfaced filaments about 10 nm in diameter (Figs. 10, 11) that were localized in a random network in the perinuclear region. Some residual actin filaments could also be observed in the peripheral regions of the cells.

After growth of smooth muscle cells in medium containing 10⁻⁶ M colchicine, the pattern of staining with the 55000 Dalton protein antibody was considerably altered. In some cells which became rounded the fluorescence was localized in a nuclear ring, while in others the staining was in the form of straight fibrils which extended the length of the cell (Fig. 8).

Isolated smooth muscle cells and cell ghosts. Smooth muscle cells isolated from the taenia coli of the guinea pig and permeated with Triton X-100 were stained directly or after extraction at high and low ionic strength. As shown previously (Small & Sobieszek, 1977 a) the latter extracted, or ghost cells are composed primarily of 10-nm filaments together with some residual actin.

With the experimental antibody, both the ghost cells and the non-extracted smooth muscle cells revealed a fine continuous network of filaments running along their entire length, with the nuclear region appearing as a sausage-shaped hole (Fig. 12). Adsorption of the antibody with smooth muscle actomyosin produced no diminution in staining (Fig. 13). However, absorption of the antibody with the purified 55000 Dalton protein (Fig. 14) gave a level of staining identical to that observed with the pre-immune serum (Fig. 15).

Cultures of rat and chick ventricle contain cardiac muscle cells, fibroblasts and some endothelial cells. Under phase-contrast the muscle cells appear irregular or polygonal. Cross-striations are often visible and the cells beat spontaneously at a rate of 30–60 beats per min.

After staining with the antibody, rat cardiac muscle cells in culture and in frozen sections exhibited two distinct patterns. In one, the antibody appeared to be localized in the region of the Z-disk (Figs. 16, 17) and was often found associated with the intercalated disk and areas of membrane between laterally associated muscle cells (Fig. 17). In other cells, however, which were seen in phase contrast to contain myofibrils and to contract spontaneously the staining was in the form of an irregular filamentous network which extended throughout the cell (Fig. 18). Adsorption of the antibody with actomyosin produced no diminution in staining, while absorption of the antibody with the purified 55000 Dalton protein resulted as before in a level of staining identical to that with the pre-immune serum (Fig. 19).

Chick cardiac muscle cells, both in culture (where they contracted spontaneously) and in frozen sections showed staining in a diffuse network throughout the cell, with no staining in the region of the Z-disk.

Myoblasts from the chicken breast muscle and rat thigh muscle begin to coalesce into groups within 3–5 h of plating, and myotubes that twitch spontaneously develop within 24 h.

When stained with the antibody, chick myotubes which showed cross-striations under phase-contrast microscopy and which twitched spontaneously, revealed a fine continuous network of filaments apparently running randomly throughout the entire cell (Fig. 20). Myoblasts also stained in a similar pattern. The same staining pattern was seen in fresh frozen sections. In the presence of colchicine the myotubes formed myosacs (see Holtzer, Sanger, Ishikawa & Strahs, 1973) which stained intensely with the antibody.

Rat myotubes, in culture and in frozen sections stain both in a diffuse network and in the region of the Z-disk.

It is particularly noteworthy that fibroblasts, normally present in cultures of smooth and cardiac muscle tissues did not stain with the antibody against the 55000 Dalton subunit even after their 10-nm filaments had been aggregated by colchicine treatment (Fig. 8).

In contrast, endothelial cells from the human umbilical cord and saphenous vein, rat ventricle and rabbit aorta stained intensely with the antibody in a fine filamentous meshwork (Fig. 21). When the cells were treated with colchicine, the filament network aggregated in a tight ring around the nucleus (Fig. 22).

After 1–2 days in culture strips of Auerbach’s plexus consist of small ganglia connected by nerve fibres and associated glial cells. Few muscle cells are present. After 5–15 days the neurons comprising the small ganglia can be clearly distinguished and many nerve fibres grow from them. Nerve fibres are visible around the expiants of sympathetic ganglia within 24 h of culture. These fibres appear as both single fine fibres and bundles of varying size often closely invested with Schwann cells. A number of non-neuronal cells are present among the nerve fibres (see Chamley et al. 1972). None of the cell types in these cultures was stained by the antibody to the 55000 Dalton polypeptide.

The presence of 10-nm filaments in vertebrate smooth muscle cells has been well documented in various electron-microscope studies (see Shoenberg & Needham, 1976). These filaments are often associated with dense bodies among the myofilament bundles or dense areas along the plasma membrane, being otherwise distributed randomly throughout the cytoplasm, sometimes singly or in small bundles, and may be observed aggregated in the juxtanuclear region of the cell (Uehara et al. 1971). They are particularly numerous in developing smooth muscle and smooth muscle cells in culture (Campbell et al. 1971). The antibody to the 55000 Dalton protein of 10-nm filaments of chicken gizzard described in this paper stained smooth muscle in a diffuse randomly organized network throughout the cell, a pattern consistent with ultrastructural studies. The staining pattern was altered but not diminished by the growth of the muscle cells in 10⁻⁶ M colchicine and the staining could be completely absorbed by the purified 10-nm filament protein, skeletin (Small & Sobieszek, 1977 a) obtained from pig stomach.

The antibody also stained 10- and 19-day embryonic chick skeletal and cardiac muscle in a diffuse, randomly organized network with an aggregation of the stained 10-nm filaments in the presence of colchicine. These findings are consistent with those of Blose et al. (1977) who raised an antibody directed against bovine brain (or glial) filaments that stained chick cardiac muscle cells. However, they differ from those of Lazarides & Hubbard (1976) who demonstrated staining with an antibody directed against the 55000 mol. wt component of chicken gizzard in the region of the Z-disk of chick skeletal muscle myofibrils. These latter authors suggested that the filaments in this region connected adjacent myofibrils to each other and to the cell membrane. In the present paper newborn rat cardiac and skeletal muscle cells stained in 2 patterns. In one, the stain was localized in a regular pattern in the region of the Z-disk and was often associated with the intercalated disk and areas of membrane between laterally associated muscle cells. The other form of staining was an irregular filamentous network which extended throughout the cell. While the reason for this discrepancy in results is at present unclear, the possibility exists that some changes in distribution of the filaments, or the protein, occur during development. In the cytoplasm of developing cardiac and skeletal muscle cells 10-nm filaments appear randomly arranged (Ishikawa et al. 1968; Rash et al. 1970), and decrease in number as cell differentiation proceeds. In more developed cells, 10-nm filaments, as well as being randomly distributed throughout the cell, have been observed in association with Z-disks (Ferrans & Roberts, 1973 a) either bridging adjacent Z-disks (Uehara et al. 1976) or passing from the Z-disk to the nuclear region. A recent study of rat heart muscle after long-term treatment with anabolic steroids, substances which stimulate protein synthesis and cell growth, has demonstrated a large number of these filaments passing between Z-disks of adjacent myofibrils (Behrendt, 1977). This arrangement at a certain stage of development could certainly account for the observed pattern of staining reported in the present paper and by Lazarides & Hubbard (1976).

On the basis of immunofluorescent staining we could identify one class of 10-nm filaments common to the 3 different muscle types in culture. Apart from differences in the patterns of staining this is consistent with the data of Lazarides & Hubbard (1977). More recently, Whalen, Butler-Browne & Gros (private communication) have shown a single type of 10-nm filament protein in all 3 muscles in adult mammals on the basis of 2-dimensional gel electrophoresis; the mammalian 50000 mol. wt skeletin subunit showed, however, only one spot with a different isoelectric point from either the a or components of gizzard (Izant & Lazarides, 1977). The absence of staining of fibroblasts present in the muscle cultures, which from electron microscopy are known to contain significant numbers of 10-nm filaments (Small & Celis, 1978b; and unpublished observations) points to the existence of a second class of 10-nm filaments. Although the data with autoimmune antibodies (Kurki et al. 1977; Osborn et al. 1977; Gordon et al. 1978) are more difficult to interpret owing to the possible heterogeneity of these sera (see Osborn et al. 1977), some obvious differences in staining have emerged. From studies of numerous human sera Kurki and coworkers (Kurki et al. 1977, and private communication) have so far distinguished immunologically distinct 10-nm filaments in cultured fibroblasts, glial cells and epidermal cells. But while these data provide evidence for the existence of several distinct classes of 10-nm filaments, a rigorous classification will be possible only from the detailed characterization of the 10-nm filament protein in the various locations in which the filaments occur.

This work was supported by the National Heart Foundation of Australia (G.R.C), the Life Insurance Medical Research Fund of Australia and New Zealand (J.C-C.), the Deutsche Forschungsgemeinschaft (U.G-S), the Volkswagen Foundation, the Muscular Dystrophy Association Inc. (J.V. S) and the Aarhus Universitets Forskningsfond (P.A).