Non-muscle myosin isoforms and cell heterogeneity in developing rabbit vascular smooth muscle

Several biochemical, immunochemical and immunocytochemical studies have demonstrated that cytoskeletal and cytocontractile proteins are valuable markers of vascular smooth muscle cell (SMC) differentiation (Gabbiani et al., 1981; Kocher et al., 1984; Kocher and Gabbiani, 1986a; Larson et al., 1984a,b; Rovner et al., 1986; Kawamoto and Adelstein, 1987; Borrione et al., 1989; Kuro-o et al., 1989; Zanellato et al., 1990a; Kuro-o et al., 1991). In the immature vessels, vimentin is the predominant component of the intermediate filament system whereas desmin content of vascular SMC increases in parallel with the progressive development of the contractile performance (Berner et al., 1981; Schmid et al., 1982; Kocher et al., 1985). The two major cytocontractile proteins, actin and myosin exist as structural variants (isoforms) whose expression is developmentally regulated (Kocher et al., 1985; Kocher and Gabiani, 1986a; Skalli et al., 1986; Borrione et al., 1989; Kuro-o et al., 1989; Zanellato et al., 1990a). An actin isoform transition from βNM-to SM-specific α-type occurs in the rat aorta concomitant with the maturation of the vascular wall (Gabbiani et al., 1984; Kocher et al., 1985; Kocher and Gabbiani, 1986a; Owens and Thompson, 1986). Concerning myosin, it has been demonstrated that the expression of 200 kDa component of myosin heavy chain isoforms (MHC) of SM-type found in the adult rabbit aorta (Eddinger and Murphy, 1988; Nagai et al., 1988,1989; Borrione et al., 1989) is preceded by a distinct isoform in the early phase of development (Borrione et al., 1989; Kuro-o et al., 1989, 1991).

Data from several laboratories have indicated the existence of one (Larson et al., 1984a,b; Borrione et al., 1989; Gaylinn et al., 1989; Shohet et al., 1989; Sartore et al., 1989) or two (Katsurawaga et al., 1989; Zanellato et al., 1990a; Kawamoto and Adelstein, 1991) myosin isoforms of NM-type in vascular SM tissue. In cultured aortic SMC obtained from adult animals, the NM myosin content can change according to the cell density and the specific phase of cell growth (Thyberg et al., 1990). The amount of NM myosin isoform in primary cultures and in subcultures from adult aorta is inversely related to that of SM myosin isoform (Rovner et al., 1986; Kawamoto and Adelstein, 1987).

Coexpression of SM and NM myosin phenotypes is especially evident in developing bovine (Zanellato et al., 1990a) and rabbit (Borrione et al., 1989; Zanellato et al., 1990b) aortic SM. In the aorta of the bovine foetus, different combinations of SM and NM myosin isoforms give rise to distinct SMC populations (Zanellato et al., 1990a). The aorta of newborn rabbits is homogeneously composed of SMC expressing SM and NM myosins, whereas very few medial cells showing this myosin isoform pattern are present in 90-day-old animal (Zanellato et al., 1990b).

Using three monoclonal anti-NM myosin antibodies we have recently demonstrated that bovine aortic endothelial cells grown in vitro show different intracellular localization of myosin immunoreactivity (Borrione et al., 1990). In the fight of this result, it would be of interest to investigate whether: (1) multiple NM are also present in rabbit aortic SM; and (2) the expression of these isoforms is developmentally regulated.

Crude myosin preparation of rabbit platelets was obtained as previously described for human platelets (Borrione et al., 1989). Rabbit platelets, thoracic aortas from 29-day-old rabbit foetuses, and from animals at days 3,15, 30, 60, 90 were used to prepare sodium dodecyl sulphate (SDS) extracts. Small specimens of aortic tissue, devoid of endothelium and adventitial tissue, were homogenized in Laemmli’s sample buffer (Laemmli, 1970) using a glass homogenizer. Samples were centrifuged at 12,000 g in a Eppendorf Minifuge (Hamburg, Germany) for 10 min and the supernatants boiled in Laemmli’s sample buffer solution for 3 min. SDS-tissue extracts were spun down at 12,000 g for 10 min and then stocked at −80°C in small samples. The apparent molecular weight of the antigen recognized by NM-F6 and NM-G2 antibodies (see Fig. 3, below) was calculated by using the high molecular weight standards of Biorad (BioRad Laboratories, Richmond,Virginia). Crude extracts from liver, spleen, kidney, intestinal epithelial cells, lung and brain of 90-day-old rabbit were prepared as follows. Small specimens from these tissues were homogenized in 0.6 M NaCl, 40 mM MOPS, 5 mM sodium-EDTA, 5 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulphonyl fluoride, 2 mg/1 leupeptin, pH 7.6, using an Ultra-Turrax homogenizer (IKA-Labortechnik, Staufen i. Br., Germany). The slurry was centrifuged at 7,000 g for 10 in a Minifuge. After adding Laemmli’s sample buffer solution, the samples were boiled for 3 min and then centrifuged twice at 16,000 g, to get rid of molecular aggregates.

Aortas from 29-day-old foetuses and from 90-day-old male New Zealand White (NZW) rabbits were used in tissue culture experiments. Primary cultures of aortic SMC were obtained by enzymic digestion of foetal and adult SM tissues following the procedure of Ives et al. (1978) with minor modifications. Collagenase (Sigma, St. Louis, MO) and elastase (Sigma) were used at 1 mg/ml and 3 mg/ml, respectively. Micro-explants obtained from this dispersion procedure were seeded on glass coverslips at density of 4 × 10³ explants/coverslip (area of coverslip: 0.48 cm²) in Dulbecco’s Modified Eagle’s Medium containing 10% foetal calf serum (Flow Laboratories, Irvine, Scotland). Tissue culture medium was replaced every 24 h. Aortic SMC in the subconfluent (after 5 days of cultivation) or confluent (after 7 days of cultivation) were rinsed in phosphate-buffered saline (PBS) and fixed in cold (—20°C) acetone for 5 min. After air drying, aortic tissue cultures from foetal and adult animals were processed for immunofluorescence as described below.

Monoclonal anti-human platelet myosin antibodies NM-G2, NM-F6 and NM-A9 are reactive specifically with platelet MHC of human origin (Borrione et al., 1989; Borrione et al., 1990; Zanellato et al., 1990a,b). SM-E7 anti-SM myosin antibody has been shown to bind selectively to MHC-1 and -2 of SM-type (Borrione et al., 1989; Zanellato et al., 1990a,b). Monoclonal anti-desmin and anti-vimentin antibodies were purchased from Boehringer-Mannheim, Germany. Monoclonal anti-SM α-actin was obatined from Sigma.

Electrophoresis was performed according to Laemmli (1970) or following the technique of Kawamoto and Adelstein (1991). In the first procedure, crude myosin preparation from rabbit platelets and crude extracts of aortic tissues from animals at the various stages of development were electro-phoresed in 10 or 4% SDS-gels. In the second prodedure, the crude extracts from the different NM tissues were electro-phoresed in 5% SDS-gels. The Western blotting technique used in these experiments was that of Towbin et al. (1979) with a few changes (described by Borrione et al., 1989; Zanellato et al., 1990a,b). The polypetides separated in the gels were blotted onto the 0.45 μm nitrocellulose paper (BioRad), and then reacted with anti-NM or anti-SM myosin antibodies. Nitrocellulose papers were then incubated with a rabbit IgG anti-mouse IgG conjugated with horseradish peroxidase (Dako, Dakopatts a/s, Glostrup, Denmark). Bound antibodies were revealed by a 3,3’-diaminobenzidine solution containing hydrogen peroxide. Parallel gels were also run to identify the position of MHC in nitrocellulose paper reacted with anti-myosin antibodies. The following controls were performed: (1) mouse non-immune IgG instead of the monoclonal anti-myosin antibodies, in the first step of the Western blotting procedure, and (2) the secondary antibody alone.

Cryosections (4 μm thick) used in this study were from: (1) ventricular myocardium and renal parenchyma of 90-day-old rabbit, and (2) thoracic aortas of 19-, 29-day-old foetuses or animals at days 3, 7, 15, 30, 45, 60 or 90 post-natal. Unfixed cryosections or cultures of aortic SMC were incubated at 37°C for 30 min with the appropriate dilution of anti-vimentin (1:8), anti-desmin (1:8), anti-SM tv-actin (1:800), SM-E7 (IgG, 5 μg/ml), NM-G2 (ascites, 1:1,500), NM-A9 (ascites 1:400) or NM-F6 (supernatant, 1:10) antibodies. After rinsing with PBS the sections/cultures were treated with a rabbit IgG anti-mouse IgG coupled with tetramethylrhoda-mine (RITC) or fluorescein (FITC) isothiocyanate (Dako), under the conditions described above. Double immunofluorescence experiments were performed using SM-E7 or NM-A9 directly coupled with FTTC and the other anti-NM myosin antibodies indirectly stained with anti-mouse IgG coupled with RITC, according to the procedures described by Zanellato et al. (1990a,b) and Borrione et al. (1990). The section/cultures were rinsed in PBS, fixed in 1.5% p- formaldehyde in PBS for 10 min at room temperature, and then mounted in Elvanol. The specimens were observed by a Zeiss Axioplan fluorescence microscope, equiped with HBO 100 W mercury lamp and Planapo optics.

Anti-vimentin and anti-desmin antibodies were applied to cryosections of vascular SM tissue from foetal, newborn and adult aortic SM tissue to test whether the intermediate filament proteins are differently distributed during development. Fig. 1 shows the immunofluorescence pattern of aortic SM tissue from 19-day-old foetuses and 3- and 90-day-old rabbits. While aortic SMC from the three developmental stages examined are homogeneously and brightly labelled with anti-vimentin antibody (Fig. 1A-C), rare SMC are stained with anti-desmin antibody (Fig. 1D-F). In the adult animal, SMC weakly reactive with anti-desmin antibody are especially evident in the third outer layer of aortic media (Fig. IF). To test whether the different levels of desmin immunofluorescence in aortic SMC may be attributable to some antigenic masking of the epitope recognized by the anti-desmin antibody, we have treated aortic cryosections with ionic (0.1%SDS in PBS, for 10 min at 25°C) and non-ionic (0.1%Triton in PBS for 10 min at 25°C) detergents as well as with some fixatives (acetone, methanol, p-formaldehyde). The desmin immunofluorescence pattern does not show any appreciable change with the various detergents and fixatives used in this experiment, except for methanol (data not shown). Thus, it is likely that in aortic SM tissue the desmin content, but not the antibody immunoreactivity, can vary during development. Anti-SM a-actin antibody when applied to cryosections from foetal, postnatal and adult aorta gave the immunofluorescence patterns shown in Fig. 2. All the aortic SMC were found to be reactive with this antibody, irrespective of the stage of development examined.

Monoclonal anti-NM myosin antibodies used in this study have been obtained utilizing actomyosin from human platelets as immunogen (Borrione et al., 1989; Borrione et al., 1990; Zanellato et al., 1990 a,b). While for NM-G2 antibody cross-reactivity with rabbit platelets has already been established (Borrione et al., 1989; Zanellato et al., 1990b), nothing is known for the other two anti-NM myosin antibodies. Immunoreactivity of NM-F6 and NM-A9 anti-human platelet MHC antibodies with rabbit platelets, as determined by Western blotting, is shown in Fig. 3 A. A component of crude myosin preparation from rabbit platelets is reactive with NM-F6 (lane 2) and NM-A9 (lane 3) antibodies. The band recognized by both antibodies shows an apparent molecular mass of about 200 kDa and comigrates with the MHC (not shown). No crossreactivity of anti-NM myosin antibodies with vascular SM myosin or tissue from adult rabbit has been demonstrated in Western blotting or in immunofluorescence experiments (for NM-G2 see Fig. 5, this study; Borrione et al., 1989; Zanellato et al., 1990b. For NM-F6 and NM-A9, see Figs 5 and 7, respectively, below).

Better detail of the specificity of anti-NM myosin antibodies can be seen in Fig. 3 B. This figure shows Western blots of the three anti-NM myosin antibodies with a number of NM tissues, electrophoresed following the procedure of Kawamoto and Adelstein (1991). In all the crude SDS extracts from spleen, liver, kidney, intestinal epithelium and lung parenchyma a single band is consistently recognized by all the three antibodies. This band has an apparent molecular mass of about 196 kDa in spleen, liver, kidney, intestinal epithelium, lung parenchyma and platelets (not shown), and of about 200 kDa in the brain extract (lane 6).

Specificity of NM-F6, NM-A9 and NM-G2 anti-NM myosin antibodies was also tested on cryosections from cardiac muscle and renal parenchyma (Fig. 4). While NM-F6 and NM-G2 stain strongly the endothelium of the capillaries present in the cardiac muscle (Fig. 4 A and E), NM-A9 selectively recognizes the pericytes (Fig. 4C). Further differences in patterns of myosin antigenicity distribution can also be observed in the renal parenchyma (Fig. 4 B,D and F). The interstitial cells are labelled with all the three antibodies, though heterogeneously with NM-A9 (Fig. 4D). In contrast, the renal corpuscles are negative with NM-G2 (Fig. 4F), positive with NM-F6 (Fig. 4B), and heterogeneous with NM-A9 (Fig. 4D). These results indicate that the anti-NM myosin antibodies display different patterns of immunoreactivity and, thus, point to the existence of multiple NM myosin isoforms in NM tissues.

Immunoreactivity of anti-SM and anti-NM myosin antibodies with SDS extracts from aortic tissues of rabbits at different stages of development was examined by Western blotting. SM-E7 antibody does not show any appreciable reactivity with NM myosin or tissue when tested by Western blotting or immunofluorescence experiments (Borrione et al., 1989; Zanellato et al., 1990a,b). Electrophoresis of aortic SDS extracts from foetal, post-natal and adult rabbits, carried out according to Laemmli (1970), shows the presence of two components in the MHC region (Fig. 5, line a); the slower migrating band corresponds to SM-MHC-1 isoform (205 kDa apparent molecular mass; Borrione et al., 1989), whereas the faster migrating band contains the SM-MHC-2 and the NM-MHC isoforms (200 kDa apparent molecular mass; Borrione et al., 1989). Incubation of nitrocellulose paper strip with SM-E7 anti-SM myosin antibody reveals that foetal and early post-natal extracts contain the SM-MHC-1 component, whereas the SM-MHC-2 becomes detectable at day 15. At days 30 and 60 this last isoform is expressed at the same relative amount with respect to SM-MHC-1, as in the aorta from 90-day-old rabbit (Borrione et al., 1989; not shown). Thus, a molecular transition at the level of SM-MHC-2 takes place during development in the aortic SM tissue. NM-F6 and NM-A9 anti-NM myosin antibodies recognize a single band in the extract from foetal aorta (lines c and d). Conversely, NM-G2 (line e) is able to react with a band present in the aortic extract from foetal as well as from the animals at days 3, 15, 30 and 60. Incubation of nitrocellulose paper strips corresponding to those shown Ln Fig. 5 (Unes c-e) with SM-E7 antibody has revealed that the bands recognized by the anti-NM myosin antibodies are localized at the level of the faster migrating component of the MHC doublet (line a; not shown). No immunoreaction is visible with the aortic preparation from 90-day-old rabbits. The results of this experiment are compatible with the presence in the rabbit aorta of at least two NM-MHC isoforms during development.

The SM and NM myosin isoform distribution in the early and in the late phase of development was analyzed by indirect and double immunofluorescence assays on aortic cryosections. Fig. 6 shows the immunostaining pattern of aortic SMC from 19-day-old rabbit foetus treated with NM-G2 and SM-E7 antibodies in a double immunofluorescence experiment. While NM-G2 antibody stains the whole vascular wall (Fig. 6A), SM-E7 labels only the innermost region of the foetal aorta (Fig. 6B). NM-F6 and NM-A9 antibodies show a pattern of myosin distribution in the aortic wall similar to that of NM-G2 (see Fig. 7A; and data not shown). In 29-day-old foetus, cellular heterogeneity in aorta with respect to SM myosin distribution has disappeared (Fig. 6D).

Concerning NM myosin isoform content in aortic SMC during development, Figs 7 and 8 show the immunostaining pattern of aortic SM cryosections taken from early foetus to adult animal with NM-F6 and NM-G2 anti-NM myosin antibody. Aortic SMC from 19- and 29-day-old foetuses are stained by NM-F6 (Fig. 7), whereas no immunoreactivity can be demonstrated with SMC of aortic media in day 3, post-natal and 90-day-old rabbit (Fig. 7 C and D). Similar results are obtained with NM-A9 (data not shown). Fig. 8 shows the immunofluorescence pattern of NM-G2 on aorta from foetal, neonatal and adult rabbit. While strong and homogeneous immunostaining is visible in the aortic SM from foetal and early postnatal animals (Fig. 8 A-C), a minority of SMC are recognized by NM-G2 in aorta from 45-day-old rabbit (Fig. 8D). Even fewer NM-G2-positive cells are present in aorta from 90-day-old rabbit (Fig. 8 E). To test whether the differental pattern of NM myosin immunostaining displayed by NM-F6 and NM-A9 on the one hand, and NM-G2 on the other, is due to some masking of the antigenic epitopes, we have treated the cryosections for 10 min at 37°C with ionic and nonionic denaturants (Zanellato et al., 1991). There is no change in the intensity and distribution of immunofluorescence in the aorta after this treatment (results not shown). The results of the immunofluorecence experiments in the rabbit aorta show different levels of SMC heterogeneity with development: (1) in the early foetus, two concentric aortic regions with distinct SM myosin content, and (2) a three-step cellular transition concerning changes in the NM myosin distribution.

The NM myosin immunoreactivity of cultured aortic SMC from foetal (29-day-old) and adult (90-day-old) rabbit is presented in Fig. 9. In both types of cultures, the distribution of NM-F6 immunostaining in SMC cytoplasm is similar irrespective of cell density (Fig. 9 A, and data not shown), i.e. long fluorescent cables that give rise to the “stress fibre” network. In both the foetal and adult cultures, NM-A9 immunostaining appears to be of punctate-type and diffuse throughout the whole cytoplasm, with almost no evidence of a distribution of stress fibre-type (Fig. 9 B). There was no difference in the appearance of NM-A9 immunostaining when cultures in a pre- or a confluent state of growth were used (not shown). In confluent cultures of aortic SMC obtained from foetal rabbit, NM-G2 immunostaining is mainly localized at the level of cell periphery (Fig. 9C). In the subconfluent phase of growth, the majority of SMC show the typical stress-fibre arrangement with almost no evidence of a distribution in the cortical cytoplasm (Fig. 9 D). This immunostaining pattern is similar to that found with NM-F6 (Fig. 9 A). In the cultures of aortic SM tissue from adult rabbit, NM-G2 immunostaining is distributed both at the level of the stress fibre system and at the cell periphery, irrespective of the degree of confluence (Fig. 9 E and F). These data indicate that: (1) NM myosin isoforms are differently localized in the cytoplasm of cultured aortic SMC, and (2) some reorganization of the intracellular NM structure identified by NM-G2 antibody takes place during SMC growth in vitro.

The use of a panel of monoclonal antibodies specific for the major cytocontractile and cytoskeletal proteins has allowed us to define in detail the sequence of maturation events that take place during development in the rabbit aorta.

Vimentin and α-actin of SM-type are present in all the aortic SMC throughout development. Conversely, in all the developmental stages examined desmin is distributed in a heterogeneous manner among vascular SMC. Our data on vimentin and desmin distribution in the rabbit thoracic aorta are in keeping with the findings reported for rat and human systems (Gabbiani et al., 1981; Schmid et al., 1982; Kocher et al., 1984, 1985; Skalli et al., 1986; Babaev et al., 1990). In contrast, Berner et al. (1981) found that rabbit abdominal aorta contains only vimentin. This apparent discrepancy might be due to the type of anti-desmin antibody used by these authors, or to the different aortic level from which the cryosections were obtained (Osborn et al., 1981; Kocher and Gabbiani, 1986b).

Concerning the aortic SMC composition in terms of SM αactin content, Skalli et al. (1986) have described a small aortic SMC population unreactive with the anti-α-actin of SM-type in adult rat aorta. In vivo, the β-NM to α-SM actin isoform transition (Kocher et al., 1985; Kocher and Gabbiani, 1986a; Owens and Thompson, 1986) is accompanied by a concomitant increase in size of the vascular SMC population containing α-actin of SM-type (Skalli et al., 1986). In the thoracic segments of developing and adult rabbit aorta, we have not been able to find any cell negative for this cytocontractile protein. Though α-SM actin can be used as a good marker of SMC origin (Darby et al., 1990), our data would indicate that in the rabbit system this protein is not a suitable marker of vascular SM differentiation.

We have obtained several pieces of evidence that indicate that multiple NM myosin isoforms exist in vascular SM tissue. These data rely on the fact that the three anti-human platelet myosin antibodies used in this study, NM-F6, NM-A9 and NM-G2, are likely to be directed against antigenic epitopes localized on different molecules. In this regard we have found that: (1) the antibodies can identify distinct microfilamentous structures in cultured endothelial (Borrione et al., 1990) and SMC cells (this study); (2) in the ventricular myocardium NM-F6 and NM-G2 on the one hand, and NM-A9 on the other, stain the endothelium and the capillary pericytes differently; and (3) in the renal parenchyma, NM-F6 and NM-A9 can label the corpuscles, whereas NM-G2 is unreactive. Selective masking of antigenic epitopes cannot account for the observed differences in the immunostaining patterns, since pre-treatment of cryosections with ionic and nonionic denaturants does not change the immunoreactivity of aortic SMC with the three antibodies.

A number of studies performed at the protein and gene levels have shown that the vascular SM coexpress NM and SM myosin isoforms (Larson et al., 1984a; Borrione et al., 1989; Sartore et al., 1989; Gaylinn et al., 1989; Shohet et al., 1989; Zanellato et al., 1990a,b; Kawamoto et al., 1991). More recently, two NM-MHC isoforms, designated as MHC-A and MHC-B (Kawamoto and Adelstein, 1991), have been found in the chicken aorta as well as in rat nonmuscle and muscle tissues and in some human cell Unes (Simons et al., 1991). Our data indicate that a third NM myosin isoform is likely to be present in the rabbit aortic SM tissue and in cultured SMC.

Besides the vascular SMC, there are other cell types that show transiently or permanently coexpression of SM and NM cytocontractile proteins, i.e. the myofibroblasts (Darby et al., 1990) and the pericytes. These latter cells, which are capable of regulating the capillary blood flow, contain NM myosin (Joyce et al., 1985) and actin (Herman and D’Amore, 1985) along with SM myosin (Joyce et al., 1985) and α-actin of SM-type (Skalli et al., 1989). Even though pericytes and vascular SMC share some structural and functional characteristics, the finding that pericytes are recognized by NM-A9 exclusively, not by NM-G2 (see below), clearly indicates that these cells possess a distinct phenotype.

The functional implications of these findings is still unknown. Some information obtained from the study of the vertebrate NM cells may be instrumental in interpreting the role of NM myosins in the vascular SM system. In the NM cell system, at least two NM-MHC isoforms are expressed in different tissues or cells (Burridge and Bray, 1975; Peleg et al., 1983, 1989; Groschel-Stewart et al., 1985; Saez et al., 1990; Murakami et al., 1990; Kawamoto and Adelstein, 1991; Simons et al., 1991). The recent discovery of an actin-based motor myosin I (Korn and Hammer, 1988; Kiehart, 1990; Pollard et al., 1991) in the brush border of the mammalian epithelia associated with a membrane fraction raises the question of whether essential cellular functions, such as cell movement or changes in the cell shape, are carried out by distinct myosin molecules (Spudich, 1989). It is interesting in this respect that cultured aortic SMC from foetal rabbit (this study) and bovine endothelial cells grown in vitro (Borrione et al., 1990) show a NM myosin isoform localized at the level of the cortical cytoplasm, in the same cellular region that contains the membrane-cytoskeletal interface enriched with the β-NM isoform of actin (Hoock et al., 1991).

The second important finding reported in this paper is related to the expression of NM myosin isoforms in differentiating rabbit vascular SMC in vivo and in vitro. Three developmetal stages could be identified on the basis of a differential immunostaining of aortic SM with anti-myosin antibodies. The first phase, characterized by the triple immunoreactivity of aortic SM with NM-F6, NM-A9 and NM-G2 antibodies begins in uterus until day 3, post-natal. In the second one, which begins after birth, aortic SMC become unreactive with NM-F6 and NM-A9 antibodies, but the immunostaining with NM-G2 is still maintained. However, heterogeneity in the NM-G2 immunostaining becomes evident around day 45; medial SMC reactive with NM-G2 decrease further from day 45 to day 90. The third phase occurs around day 90 and is characterized by the absence of imunoreactivity with NM-G2 in the majority of SMC. All aortic SMC of 90-day-old rabbit contain SM myosin (Zanellato et al., 1990bj. Rare NM-G2-positive medial SMC, however, persist in aorta at this stage of development. Thus, two distinct SMC populations exist in aorta from adult rabbit, i.e. the one that undergoes a NM myosin isoform transition through temporally distinct developmental stages towards a cell phenotype containing only SM myosin, and the one characterized by co-expression of NM and SM myosin isoforms throughout the post-natal and the adult phase of development.

Changes in myosin isoform expression during development also occur with SM myosin. In agreement with data of Kuro-o et al. (1989), we have found that a SM myosin isoform transition at the level of the SM-MHC-2 (Fig. 5) takes place between day 15 and day 30 after birth. More recently, the same authors (Kuro-o et al., 1991) have described a cDNA clone specific for the “embryonic” myosin isoform present in the rabbit aorta, and the binding properties of a polyclonal antibody against the carboxyl-terminal sequence of this protein. Their data indicate that the “embryonic” isoform of 200 kDa is actually a myosin of NM-type identical with brain MHC. Thus the appearance of SM-MHC-2 in the post-natal stage of development is preceded by the presence of a “brain”-type myosin isoform in the earlier phase. The pattern of immunoreactivity displayed by NM-G2 in Western blotting and in immunofluorescence, in both normal aorta (this study) and experimental atherogenesis (Zannelato et al., 1990b), seems to be similar to that obtained by Kuro-o et al. (1991) with the antibody against the embryonic isoform.

Regional differences in the distribution of SM (actually SM-MHC-1; Borrione et al., 1989; Kuro-o et al., 1989, 1991) and NM myosin isoforms become evident in the verly early phase of development. The aortic wall consists of two structural distinct layers of cells, i.e. the innermost layer, which shows a double myosin content (SM+ NM), and the outermost layer, which contains NM myosin exclusively. These results are in keeping with the hypothesis of a progressive accumulation and organizazion of mesenchymal cells through a “coating” process around the endothelial tubes (Pardanaud et al., 1987; Schwartz et al., 1990). These mesenchymal cells would soon become true vascular SMC, thus determining the formation of a gradient of cellular maturation in which the innermost layer of developing aorta is made of muscular cells and the outermost layer still consists of undifferentiated mesenchymal cells. This maturation process is com-pleted at around day 29 in uterus, when the whole aortic wall is composed of a homogeneous population of vascular SMC with the typical double myosin content (SM +NM)

In culture, aortic SMC from foetal or adult rabbit are able to express the same pattern of NM myosin immunoreactivity. This is surprising, since the NM myosin isoform compositions of the foetal and adult SM tissues from which the cultures have been prepared are completely different in vivo. It can be suggested that when the adult aortic SMC are grown in vitro, an undifferentiated (“foetal”?) state can be re-established, possibly due to the instability of NM myosin isoform expression. This is consistent with changes (the so-called “phenotypic modulation”; Chamley-Campbell et al., 1979; Thyberg et al., 1990) in the morphological, biochemical and immunological appearance of vascular SMC when cultured in vitro. This phenomenon could also be explained as being “a de-differentiation” process (Kocher et al., 1985), in which fully differentiated SMC can recapitulate an undifferentiated state, probably because of inadequate culture conditions.

Expression of NM myosin isoform recognized by NM-G2 antibody can be modulated in vivo by the level of cholesterolaemia (Zanellato et al., 1990b) and of thyroid hormone (Giuriato et al., 1991). Intimal thickening, which occurs in these two pathological conditions (Thomas and Kim, 1987; Schwartz et al., 1986; Ross, 1986; Giuriato et al., 1991), is characterized by proliferation and migration of medial SMC into intima. The intimal SMC contain predominantly the NM myosin isoform identified by NM-G2 antibody (Zanellato et al., 1990b). It would be of interest to determine whether the intimal SMC display a “foetal” phenotype similar to that found in culture, i.e. multiple reactivity with all the anti-NM myosin antibodies.

Future studies will address: (1) the problem of the localization of the antigenic epitopes on the NM myosins, with the aim of analyzing in better detail the structural differences among the NM myosin isoforms and of elucidating the mechanism by which the different myosin isoforms are produced in vascular SMC; and (2) the potential role of the NM myosins in the proliferara-tion/migration process that occurs in vascular pathology.

This work has been supported by grants from Italian Boehringer Ingelheim and Bayropharm Italia s.r.l.. Wethank Miss Manuela Zanoni and Mr. Maurizio Moretto for their technical help. Angela Chiavegato and Marta Scatena are fellows of Bayropharm Italia s.r.l.