Factors Influencing the in Vascular Endothelial Expression of Stress Fibers Cells In Situ

The organization of actin and myosin in vascular endothelial cells in situ was studied by immunofluorescence microscopy. Examination of perfusion-fixed, whole mounts of normal mouse and rat descending thoracic aorta revealed the presence of axially oriented stress fibers containing both actin and myosin within the endothelial cells. In both species, the proportion of cells containing stress fibers varied from region to region within the same vessel. Some endothelial cells in mouse mesenteric vein and in rat inferior vena cava also contained stress fibers. Quantitative studies of the proportion of endothelial cells containing stress fibers in the descending thoracic aorta of age-matched normotensive and spontaneously hypertensive rats revealed significant differences. When animals of the same sex of the two strains were compared, the proportion was approximately two times greater in the spontaneously hypertensive rats. The proportion of endothelial cells containing stress fibers was about two times greater in males than in females of both strains. These observations suggest that multiple factors, including anatomical, sex, and hemodynamic differences, influence the organization of the endothelial cell cytoskeleton in situ. Stress fibers in tissue-cultured cells are linear, phase-dense, cytoplasmic fibrils that are demonstrable by light microscopy and consist of bundles of microfilaments (3). Immunofluorescence techniques have revealed actin and myosin, as well as accessory contractile proteins, within these structures (15, 24, 25, 40), but their functional significance in vitro has remained unclear (4). Furthermore, on only a few occasions have similar structures been observed in situ (6, 32, 42-44). Using immunofluorescence microscopy, work in this laboratory (6) demonstrated stress fibers in the scleroblasts located near the edge or over the radial ridge of the fish scale. Inasmuch as these regions are likely to experience great shearing forces, this suggested that in situ stress fibers play a role in cellular adhesion. Stress fiber-like structures were also observed within other cells but their function was interpreted as being different from that of the fibers found in the scleroblasts. As early as 1953, Palade (29) observed filaments within the basal cytoplasm of capillary endothelial cells. Subsequently, many investigators, using transmission electron microscopy, reported the presence of micro filaments (presumably containing actin; see reference 11) in the endothelial cells of various blood vessels in normotensive animals (20). In certain endothelial cells, these microfilaments were organized into bundles (7). Cross-striated microfilament bundles (simi416 lar in appearance to the stress fibers of certain tissue-cultured cells) have been observed in the arterial endothelium (aorta and cerebral arteries) of hypertensive rats (17, 19). However, similar structures were not found in normotensive rats; thus, it was postulated that cross-striated microfllament bundles are an adaptive response of the endothelial cell to hypertension (17). Similar cross-striated microfilament bundles were found in certain fibroblasts Cmyofibroblasts") present in healing wounds (18). On the basis of their structural resemblance to myofibrils, these stress fiber-like structures were postulated to be contractile (for a review, see reference 20). We chose vascular endothelium as a model system for investigating the functional significance of stress fibers in situ. Not only is there a large body of published reports on endothelial cell biology, but this tissue offers unique advantages for such a study. For example, hemodynamic forces might be expected to influence the cytoskeleton of the endothelial cell, inasmuch as such forces have been shown to influence endothelial cell shape both in vivo and in vitro (9, 33, 37). Furthermore, the availability of spontaneously hypertensive rats (28) permits the study of endothelial cells in an animal model in which both the genetic and hemodynamic factors are relatively well defined (13, 30, 39, 45). The present study was undertaken to determine whether THE JOURNAL OF CELL BIOLOGY • VOLUME 97 AUGUS~r 1983 416-424 © The Rockefeller University Press . 0021-9525/83/08/0416/09 $1.00 on July 8, 2017 jcb.rress.org D ow nladed fom only the arterial endothelial cells of hypertensive animals contain stress fibers and whether there are factors that influence the expression c~f this structure within the endothelial cell. The aorta and other large blood vessels can be fixed by perfusion in situ, and their endothelial lining can be examined as a whole-mount preparation. By this approach, large numbers of endothelial cells can be examined in their native position. Using immunofluorescence microscopy with antibodies to actin and to myosin, we demonstrated the presence of stress fibers within the endothelial cells of normal mouse and rat thoracic aorta. Furthermore, we noted clear differences between the cytoskeletal organization of arterial endothelial cells in normotensive and spontaneously hypertensive rats. Our observations suggest that the expression of stress fibers in the vascular endothelium is influenced by multiple factors, including anatomical location, sex, genetic make-up, and hemodynamic forces. MATERIALS AND METHODS Animals : Normal adult male and female CD-I mice weighing 35--45 g were purchased from Charles River Breeding Laboratories (Wilmington, MA). Normal adult female Sprague-Dawley rats weighing 125-140 g were obtained from the same source. Eight-week-old male and female Wistar-Kyoto rats (normotensive control strain) and spontaneously hypertensive rats (OkamotoAoki type [28]) weighing 140-150 g were purchased from Taconic Farms (Germantown, NY). The females of both strains were routinely ovariectomized by the supplier and were used within 1 wk of the surgery. All animals were maintained on a diet of Purina rat laboratory chow (Ralston Purina Co., St. Louis, MO) and tap water ad libitum. Perfusion Fixation: The perfusion fixation procedure of Forssmann et al. (12) was followed. The animals were anesthesized by an intraperitoneal injection of either tribromoethanol (0.2 ml of a 2.1% solution/10 g of body weight) or chloral hydrate (0.5 ml of a 7.4% solution/100 g of body weight), the abdominal aorta was cannulated, and the inferior vena cava (infrarenal or hepatic region) was cut. Following a brief (5 s) perfusion with "rinse solution ~ (0.9% NaCI, 2.5% polyvinylpyrrolidine, Mr 40,000 [Sigma Chemical Co. St. Louis, MO], 0.025% heparin [Sigma Chemical Co.], 0.5% procaine-HC1 [Sigma Chemical Co.], pH 7.4), 100-200 ml of a fixative mixture consisting of 2% formaldehyde, 0.1% picric acid, 50 mM sodium eacodylate (pH 7.4) was introduced. All animals were perfused at 90 mm Hg hydrostatic pressure for 15 min to ensure controlled distension of the blood vessel wall during ftxation. Exclusion of the rinse solution made perfusion more difficult, and the quality of the immunofluorescent image was poor, but stress fibers nevertheless were clearly visible. Both anesthetics yielded similar fixation quality and cytoskeletal staining patterns. However, as will be discussed in Results, the concentration of formaldehyde significantly influenced the cytoskeletal staining patterns obtained. After the perfusion, the descending thoracic aorta (from the lower arch to the diaphragm) was excised, pinned down on dental wax, and flooded with phosphate-buffered saline (0.85% NaCI, 10 mM sodium phosphate, pH 7.4 [PBS]). The adventitia was carefully removed, and the vessel was cut open lengthwise between the intercostal arteries (i.e., along its dorsal aspect). The original direction of blood flow was noted. The vessel was cut crosswise into multiple segments 2-3 mm in width, and each segment was processed for immunofluorescence microscopy as described below. No attempt was made to strip off the intimal lining or to otherwise make an endothelial "Hiutchen" preparation (34). The mesenterlc vein and inferior vena cava (infrarenal portion) were handled in a similar manner. Fig. 1 illustrates some of these procedures. Antibodies: Antibodies were raised in rabbits and characterized as previously described. The anti-myosin antibody is directed against human platelet myosin (15). This antibody has also been conjugated to tetramethylrhodamine and was used in the double-label experiment. The anti-actin antibody is directed against fish skeletal muscle actin and was affinity purified (6). The antitubulin antibody is directed against vinblastine-induced tubulin crystals from sea urchin eggs (16). Rhodamine-labeled goat anti-rabbit IgG was purchased from Miles-Yeda, Rehovot, Israel (lot 17179), and fluorescein-labeled sheep anti-rabbit IgG was purchased from Wellcome Reagents, Ltd., Beckerham, England (lot K5592). Immuno f l uo rescence Procedures: After dissection, the fixed aortic segments were washed three times with PBS (5 min per wash), permeabilized by treatment with -20"C acetone for 5 rain, and washed three times with PBS ( 1 min per wash) to remove the acetone. The segments then were incubated at FIGURE 1 Preparation of rat thoracic aorta segments. (a) After in situ perfusion fixation, the aorta was excised and pinned out on dental wax with the direction of blood flow indicated. Several intercostal arteries can be seen (arrowheads). (b) The adventitia was carefully dissected away, and the vessel was incised along its dorsal aspect between the paired openings (arrows) of the intercostal arteries and pinned down. (c) Serial segments, 2-3 mm wide, were cut and prepared for immunofluorescence microscopy, and their original axial position and flow orientation were noted. Bar, 10 mm. 37"C with 50 #l/segment of affinity-purified anti-actin (50 #g/ml) or antimyosin (anti-serum diluted 100 times in PBS) in a moist atmosphere for 30--45 min. After three 5-min washings with PBS, the labeled secondary antibody (diluted 100 times in PBS) was applied, and a comparable incubation (37"C for 30-45 min) was performed in darkness. For the double-label experiment, the first incubation was with anti-actin, the second with fluorescein-labeled sheep antirabbit IgG, and the third (37"C for 30-45 rain) with rhodamine-labeled antimyosin (120 #g/ml; three dyes/IgG). After three 5-rain washings in PBS, the aortic segments were mounted in 90% glycerol in PBS. For each segment, its axial position and its orientation with respect to the in vivo direction of blood flow were noted. Several control stainings were carried out to establish the specificity of the endothelial cell staining patterns. For anti-myosin, preimmune serum and antigen-absorbed immune serum were used as the primary antibody instead of the specific antibody. In the case of anti-actin, the antigen-absorbed immune IgG preparation was tested. In these controls, there was only a faint general fluorescence over the entire specimen. The vessel segments also were tested for autofluorescence as well as for the nonspecific binding of the fluorescein-labeled sheep anti-rabbit and the rhodamine-labeled goat anti-rabbit antibodies. Lowintensity autofluorescence, as well as some nonspecific binding of the labeled secondary antibodies, was noted; however, these did not interfere with the visualization of the specific staining patterns. Staining the endothelium with an anti-smooth muscle myosin antibody (31) that does not cross-react with endothelial myosin (23) gave only a faint fluorescence in the plane of the endothelium. Fluorescence microscopy was performed with a Leitz Orthoplan microscope equipped with a Ploemopak 2 illuminator with a Leitz L-2 filter block, and a WHITE ET AL. Stress Fiber Expression in Vascular Endothelial Cells 417 on July 8, 2017 jcb.rress.org D ow nladed fom Zeiss 63 x Planapo phase-contrast objective lens (NA 1.4, oil). Fluorescent images were photographed with a Leitz Orthomat automatic camera with TriX film (Eastman Kodak Co., Rochester, N-Y). The film was exposed at ASA 1000 and developed with Acufine developer (Acutine, Inc., Chicago, IL). A scale with 10-t~m divisions was photographed at the same primary magnification to calibrate the micrographs. FIGURE 2 Region of descending thoracic aorta examined. The superimposed rectangle indicates the region of the ventral aortic wall in which the majority of the microscopic observations were made. For quantitative, comparative studies in normotensive and hypertensive rats, six randomly selected areas within the rectangle were examined. Bar, I mm. Quantitative Methods: For the quantitative study comparing the proportion of aortic endothelial cells containing stress fibers, 12 Wistar-Kyoto rats (6 males, 6 females), and 12 spontaneously hypertensive rats (6 males, 6 females) were used. In each animal, 1,000-1,500 endothelial cells were examined in six randomly selected, 150 x 250-#m areas located along the ventral aspect of the thoracic aorta at the levels of the ostia of the first through sixth intercostal arteries (Fig. 2). All counts were made directly on the epifluorescence microscope and were verified independently by a second observer. Length (>4 #m) and the appearance of striations with the antimyosin antibody were the two main criteria used to identify a stress fiber.