JOURNAL OF CELLULAR PHYSIOLOGY 143:140-149 (1990)
Expression of Simple Epithelial Cytokeratins in Bovine Pulmonary Microvascular Endothelial Cells WAYNE F. PATTON, MIN UNC YOON, J.STEVEN ALEXANDER, NANCY CHUNC-WELCH, HERBERT B. HECHTMAN, AND D A V I D SHEPRO" Biological Science Center, Boston University, Boston, Massachusetts 022 75 (W.F.P., M.U.Y., I.S.A., N.C.-W., D.S.); and Harvard Medical School, Boston, Massachusetts 02 I60 (H.B.H.)
Polypeptides of bovine aortic, pulmonary artery, and pulmonary microvascular endothelial cells, as well as vascular smooth muscle cells and retinal pericytes were evaluated by two-dimensional gel electrophoresis. The principal cytoskeleta1 proteins in all of these cell types were actin, vimentin, tropomyosin, and tubulin. Cultured pulmonary microvascular endothelial cells also expressed 12 unique polypeptides including a 41 kd acidic type I and two isoforms of a 52 kd basic type I I simple epithelial cytokeratin. Pulmonary microvascular endothelial cell expression of the simple epithelial cytokeratins was maintained in culture in the presence or absence of retinal-derived growth factor, and regardless of whether cells were cultured on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic. Cytokeratin expression was maintained through at least 50 population doubling in culture. The expression of cytokeratins was found to be regulated by cell density. Pulmonary microvascular endothelial cells seeded at 2.5 x lo5 cells/cm' (confluentseeding)expressed 3.5 times more cytokeratins than cells seeded at 1.25 x l o 4 cells/cm' (sparseseeding). Vimentin expression was not altered by cell density. By indirect immunofluorescence microscopy it was determined that the cytokeratins were distributed cytoplasmically at subconfluent cell densities but that cytokeratin 19 sometimes localized at regions of cell-cell contact after cells reached confluence. Vimentin had a cytoplasmic distribution regardless of cell density. These results suggest that pulmonary microvascular endothelial cells have a distinctive cytoskeleton that may provide them with functionally unique properties when compared with endothelial cells derived from the macrovasculature. In conjunction with conventional endothelial cell markers, the presence of simple epithelial cytokeratins may be an important biochemical criterion for identifying pulmonary microvascular endothelial cells. Embryologically, epithelial cells are derived from ectoderm, endoderm, and mesoderm. Endothelial cells are mesodermally derived simple squamous epithelial cells that line the heart, blood vascular, and lymphatic channels. Until recently, i t was thought that endothelial cells only contained the intermediate filament protein, vimentin. While in general this appears to be true, a n increasing number of exceptions have been found in microvascular endothelial cells. Desmin was demonstrated immunocytochemically to be the sole intermediate filament protein in chicken capillary endothelial cells including those of the renal cortex peritubular capillary and the hepatic sinusoid (Fujimoto and Singer, 1986). The chicken exocrine pancreas capillary was found to contain both vimentin and desmin (Fujimot0 and Singer, 1986). It has also been determined by immunocytochemistry that vimentin and desmin are coexpressed in the high endothelium venules of lymph nodes in rat but not in man (Toccanier-Pelte et al., 1987). 6) 1990 WILEY-LISS, INC.
The characteristic intermediate filament components of many epithelial cells are the cytokeratins, a multigene family of 19 or more polypeptides (Moll et al., 1982). They range in molecular weight from 40 t o 65 kd and in isoelectric point from 5.0 to 8.5. Two subfamilies of cytokeratins are distinguishable on the basis of the isoelectric points of the polypeptides: type I (acidic) cytokeratins and type I1 (basic) cytokeratins (Franke et al., 1981; Lazarides, 1982; Steinert et a]., 1985). Unlike vimentin and desmin, cytokeratins are obligatory heteropolymers requiring at least one member of each subfamily to form the tetramer subunit necessary for construction of the intermediate filament (Hatzfeld and Werner, 1985). The ratio of type I to type I1 cytokeratin in the intermediate filament is 1:l (Hatzfeld and Werner, 1985). Cytokeratins are also Received October 21, 1988; accepted December 8, 1989.
*To whom reprint requests/correspondence should be addressed.
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS
classified according to their cell type of origin. Cytokeratins 1-6 and 9-17 are commonly found in stratified epithelial cells while cytokeratins 7, 8, 18, and 19 are commonly associated with simple, nonkeratinizing epithelia (Bosch et al., 1988). Cytokeratins have been detected immunologically and electrophoretically in 4-8-cell-stage embryos, morulas, and blastocysts (reviewed in Lehtonen, 1987). In the trophectoderm cells of the blastocyst they are associated with desmosomes (Lehtonen, 1987). Coexpression of cytokeratins and vimentin has been demonstrated in a variety of embryonic tissues including the splanchnopleuric mesenchyme of the rabbit embryo (Viebahn et al., 1988). The simple epithelial keratins, cytokeratins 8 and 18, were found by immunocytochemistry with vimentin in synovial and submucosal microvascular endothelial cells of man (Jahn et al., 1987). In lower vertebrates, like the African clawed toad, all endothelial cells appear to contain both vimentin and simple epithelial keratins (Jahn et al., 1987). This widespread distribution of the simple epithelial cytokeratins may indicate that they should be considered as markers of differentiated cellular function rather than as germ layer derivation markers. In this study we demonstrate that bovine pulmonary microvascular endothelial cells, unlike any of the other vascular wall cell types examined, express simple epithelial cytokeratins. Since cytokeratins have different chemical and physical properties than other intermediate filaments like vimentin and desmin, finding this protein in the pulmonary microvascular endothelial cell may be important to this cell’s performance of differentiated functions, particularly with respect to maintaining a tight interendothelial barrier.
MATERIALS AND METHODS Isolation and culturing of cells Bovine pulmonary arteries, aortas, lungs, and retinas were obtained from a local slaughterhouse. Retinal pericytes were derived from isolated bovine retinal microvessels a s described by Gitlin and D’Amore (1983). Vascular smooth muscle cells were obtained from explant outgrowths of bovine aortas according to the method of Ross (1971). Rat pulmonary microvascular endothelial cells were isolated by a previously described method (Davies et al., 1987). A cell line of r a t pulmonary microvascular endothelial cells, isolated by retrograde perfusion of microcarriers, was obtained from Dr. U. Ryan. The macrovascular endothelial cells were obtained by a scraping technique a s previously described (Chung-Welch et al., 1988; Shepro et al., 1974). Pulmonary microvessel endothelial cells were obtained from peripheral lung sections and selected for over other cell types such a s pericytes and fibroblasts by differential plating as described previously (ChungWelch e t al., 1988). All cells were cultured in DulbecCO’SModified Eagle’s medium (DME) supplemented with 10% fetal bovine serum, 0.1% penicillin, 0.1% streptomycin, 0.1% amphotericin, and 0.1% glutamine. Primary cultures were passaged a t a 1:4 split ratio and cultured in 35 mm tissue culture dishes coated with 1.5%gelatin. For the studies evaluating the effects of different extracellular matrices on keratin expression, collagen I, collagen IV, fibronectin, tissue culture plas-
141
tic, laminin, or basement membrane proteins (Matrigel) were used according to manufacturer’s suggestions (Collaborative Research, Inc., Waltham, MA, USA).
Characterization of cell types Bovine pulmonary artery endothelial cells, zortic endothelial cells, and pulmonary microvascular endothelial cells expressed the typical endothelial cell “cobblestone” morphology and stringent contact inhibition and were found to stain positively with f luorescently labeled acetylated low-density lipoprotein, factor VIII, and angiotensin-converting enzyme (Chung-Welch et al., 1988; Weinberg et al., 1982; Voyta et al., 1984). We have also previously demonstrated SQ20881-inhibitable angiotensin-converting enzyme activity in all three of the endothelial cell types using [3H] benzoyl Phe-Ala-Pro as substrate (Chung-Welch et al., 1988). SQ20881 is a n angiotensin-converting enzyme antagonist (Peninsula Laboratories, Inc., San Pedro, CAI. In addition, the cells expressed the expected prostanoid patterns, with PGI, predominant in the macrovascular endothelial cells and either PGI, or PGE, predominant in the microvascular endothelial cells, dependent upon whether or not growth factor was included in the culture media (Chung-Welch et al., 1988). By visualization of the actin cytoskeleton with rhodamine-phalloidin i t was determined that the cells were responsive to vasoactive amines such as histamine and vasoactive peptides such a s angiotensin I1 and bradykinin. Smooth muscle cells were identified by their “hill and valley” morphology, lack of contact inhibition, and lack of staining with fluorescently labeled acetylated lowdensity lipoprotein. Retinal pericytes also did not stain with fluorescently labeled acetylated low-density lipoprotein, but did not form a “hill and valley” morphology and proliferated very slowly in culture. Metabolic labeling of polypeptides For the detection of total cellular protein, the normal 10% fetal calf serum-DMEM was replaced by 90% methionine-free FCS-DMEM, 10% normal FCS-DMEM supplemented with 0.1 mCilml of Tran”5S-label (I.C.N., Irvine, CA) just prior to endothelial cells reaching confluence. Cells were then incubated for 24-48 h in the labeling media. T ~ - a n ~ ~ S - l aisb used e l a s a [35Sl-methionine substitute for metabolic labeling. It contains 709k [35Sl-methionine and 20% [35Sl-cysteine. Smooth muscle cells were also labeled for 24-48 h whereas retinal pericytes due to their slow proliferation rate were generally labeled for 72 h. Cell lysate and cytoskeleton isolation Triton-insoluble cytoskeletons were prepared by a modification of the method of Gilbert and Fulton (1985). All buffers were maintained a t 37°C during the cytoskeleton isolation. Endothelial cell monolayers in 35 mm dishes were washed twice in DMEM and then extracted for 1 min in 10 mM Hepes, 100 mM KCl, 2.5 mM MgCI,, 1 mM CaC12, 300 mM Sorbitol, 1.0 mM phenylmethylsulfonyl fluoride (PMSF), 1.0%Triton X100, adjusted to pH 7.4 (extraction buffer). The extraction buffer was carefully aspirated and the insoluble structures remaining associated with the dish were incubated for a n additional minute with fresh extraction buffer. After removing the extraction buffer by aspira-
142
PATTON ET AL.
tion, a final wash was performed using 2 ml of 10 mM Tris-C1, 1.0% Triton X-100, 1.0 mM PMSF, adjusted to pH 7.4. The Triton-insoluble cytoskeletons were solubilized by addition of 60 p1 of 95°C 2.5% sodium dodecylsulfate (SDS), 20 mM Tris-C1, 1.0 mM PMSF, 2.5% B-mercaptoethanol, adjusted to pH 8.0. By using the point of a 200 pl pipet tip, the monolayer was vigorously scraped until a n amorphous, transparent, sticky layer was released from the bottom of the dish corresponding to the cytoskeleton. In order to obtain cell lysates, monolayers were washed twice in DMEM and then solubilized by addition of 100 pl of 95°C2.5% SDS, 20 mM Tris-C1, 1.0 mM PMSF, 2.5% B-mercaptoethanol, adjusted to pH 8.0. Monolayers were then scraped in a similar manner to the Triton-insoluble cytoskeletons.
molecular weight standards (Sigma Chemical Co., St. Louis, MO). Though molecular weights and isoelectric points were determined by using external standards, the values obtained for actin and vimentin were found to be within 3% of the literature values.
Immunofluorescence microscopy For immunofluorescence studies, endothelial cells were grown to confluency on 12 mm round glass coverslips that had previously been coated with 1.5% gelatin. Cells were fixed in 3.7% phosphate-buffered formaldehyde (pH 7.4) for 15 min. The cells on coverslips were permeabilized in extraction buffer (0.5 M KC1, 1% Triton X-100, 10 mM MgCl,, 1 mg/ml Tris HC1, 17 pg/ml TSF, 0.25 mg/ml DNAse 1 in phosphate-buffered saline (PBS)) for 10 min. This solution was removed and the cells were washed three times with PBS. The Protein electroblotting and visualization of the cells were stained with monoclonal anti-cytokeratin cytokeratins and vimentin 4.62, monoclonal anti-cytokeratin 8.12 or monoclonal The cytokeratins and vimentin were detected immu- anti-vimentin 13.2 antibody. FITC-conjugated rabbit nologically by transferring proteins to nitrocellulose anti-mouse IgG was added to visualize the intermedisheets by the method of Towbin et al. (1979). Prior to ate filaments by immunofluorescence microscopy. The electroblotting, the gels were allowed to pre-equil- coverslips were washed five times in PBS, mounted in ibrate at room temperature in electroblot buffer (50 PBS/glycerol (l:l), and sealed. Cells on coverslips were mM Tris-base, 95 mM glycine, 0.005% SDS) for 30-45 illuminated for fluorescence by using a Zeiss universal min. The transfer to nitrocellulose paper was per- microscope with a n Olympus OM-25 camera attachformed for 3 h a t 250 mA constant current and 40-50 ment. volts. The blots were incubated overnight with polyRESULTS clonal antibody produced by using bovine hooves as immunogen (a general anti-keratin antibody), monoComparison of polypeptide patterns of clonal anti-cytokeratin antibody (clone no. K4.62, spe- pulmonary microvascular endothelial cells with cific for cytokeratin 19), monoclonal anti-cytokeratin other cells of the vascular wall antibody (clone K8.12, reacts with a broad range of The distribution of polypeptides in pulmonary arkeratins including cytokeratin 81, or monoclonal antivimentin antibody (clone no. VIM-13.2, specific for vi- tery, aortic, and pulmonary microvascular endothelial mentin). The blots were then incubated with alkaline- cells, a s well a s vascular smooth muscle cells and retphosphatase-conjugated rabbit anti-mouse IgG or inal pericytes was examined by two-dimensional gel rabbit anti-guinea pig IgG diluted 1:1,000, and visu- electrophoresis. All cell types examined possessed the alized by using a nitroblue tetrazolium chloride/ cytoskeletal proteins actin, vimentin, tubulin, and 5-bromo-4-chloro-3-indoylphosphate p-toluidine chro- tropomyosin a s determined by their characteristic promogen system. All antibodies and chromogens were tein mobilities reported by others (Savion et al., 1982; Blose, 1984). Actin was also identified by its comigrapurchased from Sigma Chemical Co., St. Louis, Mo. tion with rabbit skeletal muscle actin while vimentin Two-dimensional SDS polyacrylamide gel was identified by using monoclonal anti-vimentin 13.2 electrophoresis antibody. Though each cell type was observed to have Two-dimensional gel electrophoresis was performed polypeptides not possessed by the other cell types, the according to the method of O'Farrell (1975). To reduce most striking polypeptide pattern differences were obstreaking in the isoelectric focusing dimension, the tained from pulmonary microvascular endothelial samples were prepared as described by Ames and Nai- cells. Approximately 340 polypeptides were detected by kaido (1976). The second dimension was run on 10% quantitative two-dimensional gel electrophoresis of polyacrylamide slab gels and for Tran3%-1abel studies, whole cell lysates from pulmonary microvascular ensubsequently incubated for 30 min in two changes of dimethylsulfoxide, followed by a 3 h incubation in 20% 2,5-diphenyloxazole (PPO) in dimethylsulfoxide, a 1 h incubation in water, and then a 2 h incubation in 2.5% Fig. 1. Two dimensional gel electrophoretic comparison of various glycerol prior to air drying between dialysis sheets cells of the vascular wall after metabolic labeling with Tran"5S-label. (Dunbar, 1987). Radiolabeled gels were exposed to A Retinal pericytes. B: Vascular smooth muscle cells. C: Pu1mona:ry Kodak X-OMAT film by manufacturer's instructions. artery endothelial cells. D Aortic endothelial cells. E: Pulmonary endothelial cells. F: Consensus map of polypeptides Gels were analyzed by using a microcomputer-based microvascular identified in pulmonary microvascular endothelial cells. Proteins in videodensitometer as previously described (Mariash et black were found to be unique to pulmonary microvascular endotheal., 1982; Patton e t al., 1989). Isoelectric points were lial cells. The other proteins shown in the map are expressed in macdetermined by measuring the pH of 1 cm segments of rovascular endothelial cells, smooth muscle cells, and pericytes ;as well as the pulmonary microvessel endothelial cells. Many are exblank gels run in parallel with the sample while mo- pressed in different amounts between the cell types. Two of the unique lecular weights were determined from one-dimensional polypeptides corresponded to cytokeratins 8 and 19. Vimentin and the electrophoretic separations of commercially available three principal isoforms of actin (a$,?)are also identified in the map.
Fig. 1.
144
PATTON ET AL.
dothelial cells. The average abundance of each polypeptide resolved on the gels was found to be 0.29%. Twelve major polypeptides, each representing 0.21 to 1.7% of the total cellular protein, were found to be expressed exclusively by the pulmonary microvascular endothelial cells. To identify further the polypeptides expressed uniquely in the pulmonary microvascular endothelial cells, Triton X-100-resistant cytoskeletons were isolated and evaluated. Three of the unique polypeptides observed in the cell lysate patterns were enriched substantially in the cytoskeleton preparation. The resistance of these polypeptides to extraction with Triton X-100 as well as their high abundance in the cell are properties consistent with them being intermediate filament proteins (Moll et al., 1982). The molecular weights and isoelectric points of the three polypeptides were M, = 41 kd, PI = 5.1; M, = 52 kd, PI = 5.65; and M, = 52 kd, PI = 5.9. The two polypeptides with the same molecular weight are believed to represent different isoforms of the same polypeptide. The molecular weight and isoelectric point coordinates of the pulmonary microvascular endothelial cell proteins corresponded t o the coordinates of known simple epithelial cytokeratins (Wu et al., 1982). These polypeptides correspond to cytokeratins 19 and 8 when one uses the nomenclature of Moll et al. (1982). Figure 1 shows a representative comparison made between whole cell lysate polypeptide patterns of all the cell types examined as well a s the positions of actin, vimentin, and the cytokeratins. The relative abundances of the cytokeratins, actin and vimentin were determined in pulmonary microvascular endothelial cell lysates and cytoskeletons by quantitative videodensitometry. In whole cell lysates actin represented 4.5%,vimentin represented 1.2%, cytokeratin 19 represented 1.14%, and the two isoforms of cytokeratin 8 together represented 1.7% of the total amount of protein visualized on the two-dimensional gels. The experimentally determined ratio of cytokeratin 19 to cytokeratin 8 was thus found to be 1:1.5. A comparison of the nucleotide sequences of bovine keratins 8 and 19 obtained from the National Gene Bank revealed that cytokeratin 8 contains 15 methionine residues while cytokeratin 19 contains 10. Correcting for the greater number of methionine residues in cytokeratin 8 yields a cytokeratin 8 to 19 ratio of 1:1, consistent with the known requirements for intermediate filament assembly discussed in the Introduction. In l’riton-insoluble cytoskeletons, actin represented 10.9%,vimentin represented 4.3%, cytokeratin 19 represented 3.5%, and the two isoforms of cytokeratin 8 represented 5.3% of the total amount of protein visualized on the two-dimensional gels. For comparison, i t has previously been demonstrated by two-dimensional gel electrophoresis that actin represents 3.3%, vimentin 0.55%, and each of the five keratin species represent from 0.07 to 0.63% of the total amount of visualized protein in HeLa cell lysates (Bravo and Celis, 1984). Vimentin was found to be present in pulmonary microvascular endothelial cells in only 33%of the amount found in pulmonary artery endothelial cells. The total amount of intermediate filament protein in pulmonary microvascular endothelial cells was slightly greater than pulmonary artery endothelial cells.
Immunological characterization of the simple epithelial cytokeratins and vimentin Immunological analysis with polyclonal antibody raised against epidermal keratins, monoclonal antibody reactive to a broad range of keratins, and monoclonal antibody reactive to cytokeratin 19 confirmed the identification of the polypeptides as simple epithelial cytokeratins. The identity of vimentin was also confirmed by using monoclonal antibody raised against vimentin. Western blots probed with the anti-keratin polyclonal antibody showed reactivity with a number of polypeptides in both aortic endothelial cells and pulmonary microvascular endothelial cells, including some soluble, noncytoskeletal proteins, but reactivity with the 41 kd and 52 kd polypeptides (cytokeratins 19 and 8 ) was restricted to the pulmonary microvascular endothelial cells (data not shown). A monoclonal antibody reactive to a broad range of keratins (K8.12) was found to display low crossreactivity with soluble proteins. As shown in Figure 2, monoclonal antibody K8.12 selectively reacted with cytokeratin 8 and 19 i~n pulmonary microvascular endothelial cells. The monoclonal antibody to cytokeratin 19 (K4.62) primarily reacted with the 41 kd protein in pulmonary microvascular endothelial cells while vimentin was detected in both endothelial cell types. Two-dimensional gels transferred to nitrocellulose sheets and probed with monoclonal antibody K8.12 also demonstrated the presence of cytokeratins 8 and 19 in pulmonary microvascular endothelial cells (see Fig. 3). As shown in Figure 4, immunofluorescence microscopy of monolayers of cells with monoclonal anti-cytokeratin 19, monoclonal anti-cytokeratin (K8.12), and monclonal antivimentin (VIM-13.12) antibody demonstrated that both keratin and vimentin intermediate filaments existed in the pulmonary microvascular endothelial cell cytoplasm as distinct filament networks. At subconfluence the cytokeratins showed a cytoplasmic distribution, but after the cells were cultured for extended periods of time a t confluency (1-3 weeks), cytokeratin 19 was often found to be localized a t the periphery of the cell and appeared to be associated with the cell-cell junctions. Vimentin had a more diffuse distribution in the cytoplasm regardless of the state of confluenc,y. Aortic endothelial cells had a similar distribution of vimentin but the cytokeratins were not detected in the cells. Expression of the simple epithelial cytokeratins was observed to be maintained in culture in the presence or absence of retinal-derived growth factor, and regardless of whether cells were cultured on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic through a t least 50 population doublings. The levels of keratin expression were found to be largely independent of the substrate pulmonary microvascular endothelial cells were exposed to. Cells seeded on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic a t 1:4 split ratios (6.25 x lo4 cells/cm2) and allowed to reach confluency differed in the amount of expressed keratin by less than 10-15% relative to the other cytoskeletal proteins (n = 3). The levels of keratins expressed in pulmonary microvascular endothelial cells were found to be profoundly influenced by cell
CYTOKERATTNS IN PULMONARY ENDOTHELIAL CELLS
145
Fig. 2. Immunological identification of the cytokeratins in pulmonary microvascular endothelial cells. Whole cell lysates of pulmonary microvascular and aortic endothelial cells were resolved by one-dimensional gel electrophoresis and either stained with Coomassie blue dye or electrophoretically transferred to nitrocellulose for subsequent probing with antibodies as described in the Methods section. A Coomassie blue dye patterns of microvascular and aortic endothelial cell proteins. B: Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal antibody K8.12 (reactive to a variety of acidic and basic keratins). Cytokeratins 8 and 19 were labeled only in the microvascular endothelial cells. C.
Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal anti-cytokeratin antibody (clone no. K4.62, specific for cytokeratin 19). A minor second band that is less than 10% of the keratin 19 band is also labeled with the K4.62 antibody. On 2 D gels this second band has a very acidic isoelectric point (4.6) but remains unidentified. D Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal anti-vimentin antibody (clone no. VIM-13.2). The positions of vimentin (V), cytokeratin 8 (C 81, actin ( A ) ,and cytokeratin 19 (C 19) are indicated.
density. Cells seeded at 2.5 x lo5 cells/cm2 (confluent seeding) expressed 3.5 times more cytokeratins 8 and 19 than cells seeded a t 1.25 x lo4 cells/cm2 (sparse seeding) (n = 3). As the cells became progressively more confluent, the amount of keratin in their cytoplasms steadily increased.
cardial, and peritoneal cavities as well as the outer surfaces of the lungs, heart, and viscera. Human mesothelial cells have been shown to express cytokeratins 7, 8, 18, and 19 (Wu et al., 1982; Connell and Rheinwald, 1983; Kim et al., 1987). Thus far, cytokeratin 7 has not been shown to be exmessed in bovine cells (O'Guin et al., 1987). Mesothelial cells and endothelial cells are difficult to distinguish from one another by histological criteria alone (Rheinwald et al., 1987; Satoh and Prescott, 1987). Endothelial cells present a nonthrombogenic, nonadhesive surface to the blood whereas mesothelial cells perform the same function with respect to the serous fluids (Rheinwald e t al., 1987; Satoh and Prescott, 1987). Our pulmonary microvascular endothelial cells as well as mesothelial cells have numerous microvilli projections on the apical plasma membrane surface (Chung-Welch et al., 1988; Andrews and Porter, 1973; Satoh and Prescott, 1987). Both bovine pericardial mesothelial cells and bovine pulmonary microvascular endothelial cells stain positively for Factor VIII-related antigen (Satoh and Prescott, 1987; Chung-Welch et al., 1988). We demonstrated that unstimulated cultures of pulmonary microvascular endothelial cells synthesized 0.86 ng PGI,/mg protein while 1.0 pM bradykinin and
DISCUSSION In this study, cultured bovine pulmonary microvessel endothelial cells were found to express cytokeratins 8 and 19 a s well a s the intermediate filament protein vimentin. A number of epithelial cell types including hepatocytes, colonic mucosa, and urothelium contain . for a both vimentin and keratin (see Moll et ~ 1 .1982, more extensive listing). This pattern of intermediate filament protein expression, however, is exceedingly rare among cells derived from the mesodermal germ layer and thus pulmonary microvascular endothelial cells belong to a novel class of mesodermally derived cells expressing simple epithelial-type cytokeratins as well as vimentin (Wu et al., 1982). Mesothelial cells are the only other cells derived from the mesoderm known to synthesize vimentin in conjunction with simple epithelial keratins. Mesothelial cells are the flat cells contained in the single layer lining the pleural, peri-
146
PATTON ET AL
Fig. 3. Detection of cytokeratins 8 and 19 by two-dimensional gel electrophoresis followed by Western blotting and incubation with monoclonal antibody K8.12. A Two-dimensional gel of cytoskeletal proteins detected after labeling with "ran%-label. B Two-dimen-
sional Western blot after reaction with monoclonal antibody K8.12. The positions of vimentin (V), cytokeratin 8 (C €3, actin (A), and cytokeratin 19 (C 19)are indicated.
1.9 pM of the calcium ionophore A23187 caused 2.14 and 2.66 ng PGI,/mg protein to be synthesized in the cultures respectively (Chung-Welch et al., 1988). Others have demonstrated that cultured bovine pericardial mesothelial cells synthesized 2.1 ng PGI,/mg protein whereas 10 p M of bradykinin and 5 p M of the calcium ionophore A23187 caused 18.3 and 94.6 ng PGI,/mg
protein to be synthesized in the cultures respectively (Satoh and Prescott, 1987). In humans, mesothelial cells also share with endothelial cells the expression of a prominent secreted glycoprotein (M, 46 kD, PI 7.11, commonly referred to as mesosecrin (Rheinwald et al., 1987). It should be mentioned a t this point that the pulmonary microvascular
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS
Fig. 4. Immunofluorescence localization of cytokeratin and vimentin in subconfluent pulmonary microvascular and aortic endothelial cells. A Pulmonary microvascular endothelial cells labeled with antivimentin antibody (clone no. VIM-13.2). B: Aortic endothelial cells labeled with anti-vimentin antibody (clone no. VIM-13.2). C: Pulmonary microvascular endothelial cells labeled with anti-cytokeratin antibody (clone no. K4.62,specific for cytokeratin 19). D Aortic endothelial cells labeled with anti-cytokeratin antibody (clone no.
147
K4.62). Only very weak nonspecific background staining by the rhodamine-conjugated second antibody was detectable. E: Pulmonary microvascular endothelial cells labeled with anti-cytokeratin antibody (clone no. K8.12,specific for a broad spectrum of cytokeratins including cytokeratin 8 and 18). F: Aortic endothelial cells labeled with anti-cytokeratin antibody (clone no. K8.12). Only very weak nonspecific background staining by the rhodamine conjugated second antibody was detectable. Bars, 1 p,m.
148
PATTON ET AL
endothelial cells described in this study were isolated after physical removal of the lung pleura, a potential source of mesothelial cells (Chung-Welch et al., 1988). The cell cultures obtained from the dissected peripheral lung by filtering with Nitex screens and differential plating were examined by using a battery of markers to confirm their identity as endothelial cells (Chung-Welchet al., 1988).We have examined rat pulmonary microvascular endothelial cells isolated by the filtering method of Davies et al. (1987) and by the retrograde perfusion method of Ryan et al. (1982). Both of these cell lines contained an abundance of cytokeratins 8 and 18 as well as some cytokeratin 19 as determined by two-dimensional gel electrophoresis (Dodge et al., submitted). We could not detect cytokeratin 18 in bovine pulmonary microvascular endothelial cells by twodimensional gel electrophoresis of cytoskeletal proteins or by immunoreactivity with the general anti-keratin antibody. Using high-resolution two-dimensional gels we have been able to detect cytokeratin 18 in bovine pericardial mesothelial cells. The pulmonary microvascular endothelial cells can also be distinguished from mesothelial cells on the basis of their ability to form tube-like structures on basement membrane proteins or collagen I (Chung-Welch et al., 1989). The bronchial tree is a meticulously designed system of branching airways containing an abundance of epithelial cell types (Weibel, 1980). Few studies concerning the keratin patterns of these cells have been conducted due to the difficulty in obtaining pure preparations, free of contaminating cell types. Moll et al. (1982) found cytokeratins 6-8 and 17-19 in human undifferentiated carcinoma of the lung and they observed cytokeratins 5-8, 13-15, and 17-19 in human tracheal epithelium. This is consistent with the cells originating from stratified epithelia. Blobel et al. (1984)found the simple epithelial cytokeratins 7,8,18, and 19 in the alveolar regions of the human lung by microdissection techniques and immunofluorescence microscopy. They were unable to assign these species to individual cell types in this region of the lung, however. Woodcock-Mitchell et al. (1986) examined pure cultures of Type I1 alveolar pneumocytes as well as paraffin-embedded sections of normal and injured rat lung for cytokeratin species. Type I1 alveolar pneumocytes contained cytokeratins 7, 8, and 18 but no cytokeratin 19. None of the lung epithelial cell types described expressed more than trace amounts of the intermediate filament protein, vimentin. The basic structural property of lung parenchyma is the reduction of the tissue mass to a thin sheet that provides the air and blood an extensive access to one another (Weibel, 1980). The mechanical problems inherent in suspending the delicate alveolar septa while exposing the capillary to air over a large surface require a highly organized connective tissue lattice. Ventilatory movements of the thorax and diaphragm are transmitted all along the bronchial tree through the peripheral fiber lattice system to the alveoli themselves (Weibel, 1980). The capillary endothelium as a minimal cell layer constituting a barrier between blood and surrounding tissue may require an epithelial-like intermediate filament network to compensate for the unusual mechanical burden it is subjected to by virtue of its anatomical location. Cytokeratins may function
as the intracytoplasmic portion of desmosomes, providing anchorage points between individual cells and distributing mechanical stress applied to one cell over the entire tissue. In situ analysis of human and bovine lung using monoclonal antibody specific to cytokeratin 19 demonstrates that this intermediate filament protein is associated with endothelial junctions in some microvascular beds while being cytoplasmically distributed in others (Mineau-Hanske et al., submitted). In this study we demonstrate that cultured bovine pulmonary microvascular endothelial cells, unlike any of the other vascular wall cell types examined, express two simple epithelial cytokeratins: an acidic type I keratin (M, = 41 kd, PI = 5.1) and a basic type I1 keratin (M, = 52 kd, PI = 5.7-5.9). Since cytokeratins have different chemical and physical properties than other intermediate filaments like vimentin and desmin, finding this protein in the pulmonary microvascular endothelial cell may be important to this cell’s performance of differentiated functions.
ACKNOWLEDGMENTS The authors wish to acknowledge personnel from our laboratory: Andrea Dodge for the retinal pericyte cultures; Jeanne Choi for assistance with fluorescence microscopy; and G.P. Ameia Yen-Patton, Dr. Rochelle Mineau-Hanske, and Dr. Nicole Morel for useful discussions. This investigation was performed at Boston University, Biology Department, Boston, MA 02215, and was supported in part by USPHS grants HBL16714,33104, and GM24891.
LITERATURE CITED Ames, G., and Nikaido, K. (1976)Two-dimensional gel electrophoresis of membrane proteins. Biochemistry, 15:616-623. Andrews, P., and Porter, K. (1973) The ultrastructural morphology and possible functional significance of mesothelial microvilli. Anat. Rec., 177:409-426. Blobel, G., Moll, R., Franke, W., Vogt-Moykopf, I. (1984) Cytokeratins in normal lung and lung carcimonas. Virchows. Arch. l Cell Pathol. J 45:407-429. Blose, S. (1984) The endothelial cytoskeleton. In: Biology of Endothelial Cells. E.M. Jaffee, ed. Martinus, Boston, Ch. 14, pp. 141-154. Bosch, F., Leube, R., Achtstatter, T., Moll, R., and Franke, W. (1988) Expression of simple epithelial type cytokeratins in stratified epithelia as detected by immunolocalization and hybridization in situ. J . Cell Biol., 106t1635-1648. Bravo, R., and Celis, J. (1984) Catalog of HeLa cell proteins. In: TwoDimensional Gel Electrophoresis of Proteins; Methods and Applications. J. Celis and R. Bravo, eds. Academic Press, New York, Ch. 14,pp. 445-476. Chung-Welch, N., Shepro, D., Dunham, B., and Hechtman, H. (1988) Prostacyclin and prostaglandin E, secretions by bovine pulmonary microvessel endothelial cells are altered by changes in culture conditions. J. Cell. Physiol., 135t224-234. Chung-Welch, N., Patton, W., Yen-Patton, G., Hechtman, H., and Shepro, D. (1989) Phenotypic comparison between mesothelial and microvascular endothelial cell lineages using conventional endothelial cell markers, cytoskeletal protein markers and in uitro assays of angiogenic potential. Differentiation 42:44-53. Connell, N., and Rheinwald, J. (1983) Regulation of the cytoskeleton in mesothelial cells: Reversible loss of keratin and increase in vimentin during rapid growth in culture. Cell, 34t245-253. Davies, P., Smith, B., Maddalo, F., Langleben, D., Tobias, D., Fujiwara, K., and Reid, L. (1987) Characterization of lung pericytes in culture including their growth inhibition by endothelial substrate. Microvasc. Res., 33t300-314. Dodge, A., Patton, W., Yoon, M., Hechtman, H., and Shepro, D. (1989) Organ and species specific differences in intermediate filament protein profiles of cultured microvascular endothelial cells. J. Cell Sci. (submitted).
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS Dunbar, B. (1987) Two-Dimensional Electrophoresis and Immunological Techniques. Plenum Press, New York. Eichner, R., Sun, T., and Aebi, U. (1986) The role of keratin subfamilies and keratin pairs in the formation of human epidermal intermediate filaments. J. Cell Biol., 102r1767-1777. Franke, W., Schiller, D., Moll, R., Winter, S., Schmid, E., and Engelbrecht, I. (1981) Diversity of cytokeratins. Differentiation specific expression of cytokeratin polypeptides in epithelial cells and tissues. J . Mol. Biol., 153t933-959. Fujimoto, T., and Singer, S. (1986) Immunocytochemical studies of endothelial cells in vivo. I. The presence of desmin only, or of desmin plus vimentin, or vimentin only, in the endothelial cells of different capillaries of the adult chicken. J. Cell Biol., 103r27752786. Gilbert, M., and Fulton, A. (1985) The specificity and stability of the triton-extracted cytoskeletal framework of gerbil fibroma cells. J. Cell Sci., 73:335-415. Gitlin, J., and D’Amore, P. (1983) Culture of retinal capillary cells using selective growth media. Microvasc. Res., 26t74-80. Hatzfeld, M., and Franke, W. (1985) Pair formation and promiscuity of cytokeratins: Formation in vitro of heterotypic complexes and intermediate-sized filaments by homologous and heterologous recombinations of purified polypeptides. J. Cell Biol., 1OIt1826-1841. Jahn, L., Fouquet, B., Robe, K., and Franke, W. (1987) Cytokeratins in certain endothelial and smooth muscle cells of two taxonomically distant vertebrate species, Xenopus laevis and man. Differentiation, 36:234-254. Kim, K., Stellmach, V., Javors, J., and Fuchs, E. (1987) Regulation of human mesothelial cell differentiation: Opposing roles of retinoids and epidermal growth factor in the expression of intermediate filament proteins. J. Cell Biol., 105t3039-3051. Lazarides, E. (1982) Intermediate filaments: A chemically heterogeneous, developmentally regulated class of proteins. Annu. Rev. Biochem., 51t219-250. Lehtonen, E. (1987) Cytokeratins in oocytes and preimplantation embryos in the mouse. In: Current Topics in Developmental Biology, Volume 22: The Molecular and Developmental Biology of Keratins. R. Sawyer, ed. Academic Press, New York, Chapter 7, pp. 153-173. Mariash, C., Seelig, S., and Oppenheimer, S. (1982) A rapid, inexpensive, quantitative technique for the analysis of two-dimensional electrophoretograms. Anal. Biochem., 12It388-394. Mineau-Hanschke, R., Patton, W., Hechtman, H., and Shepro, D. (1989) Immunolocalization of cytokeratin-19 in bovine and human pulmonary microvessel endothelial cells in situ. In Vitro (submitted). Moll, R., Franke, W., and Schiller, D. (1982) The catalog of human cutokeratins: Patterns of expression in normal epithelial, tumors and cultured cells. Cell, 31:ll-24. O’Farrell, P. (1975) High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem., 25Ot4007-4021. O’Guin, M., Galvin, S., Schermer, A,, and Sun, T. (1987) Patterns of keratin expression define distinct pathways of epithelial development and differentiation. In: Current Topics in Developmental Biology, Volume 22: The Molecular and Developmental Biology of
149
Keratins. R. Sawyer, ed. Academic Press, New York, Chapter 5, pp. 97-125. Patton, W., Dhanak, M., and Jacobson, B. (1989) Identification of dictyostelium discoideum plasma membrane proteins by cell surface labeling and quantitative two-dimensional gel electrophoresis. Anal. Biochem., 179t37-49. Rheinwald, J., Jorgensen, J., Hahn, W., Terpstra, A,, O’Connell, T., and Plummer, K. (1987) Mesosecrin: A secreted glycoprotein produced in abundance by human mesothelial, endothelial and kidney epithelial cells in culture. J. Cell Biol., 104:263-275. Ross, R. (1971) The smooth muscle cell 11. Growth of smooth muscle in culture and formation of elastic fibers. J. Cell Biol., 50t172-186. Ryan, U., Clements, E., Habliston, D., and Ryan, J . (1982) Use of microcarriers to isolate and culture pulmonary microvascular endothelium. Tissue Cell, 16.577-691. Satoh, K., and Prescott, S. (1987) Culture of mesothelial cells from bovine pericardium and characterization of their arachidonate metabolism. Biochim. Biophys. Acta, 930:283-296. Savion, N., Vlodavsky, G., Greenburg, G . , and Gospodarowicz, D. (1982) Synthesis and distribution of cytoskeletal elements in endothelial cells as a function of cell growth and organization. J. Cell. Physiol., 110t129-141. Shepro, D., Rosenthal, M., Batbouta, L., Roblee, L., and Belamarich, F. (1974) The cultivation of aortic endothelium. Anat. Rec., 78:523. Steinert, P., Steven, A., and Roop, D. (1985) The molecular biology of intermediate filaments. Cell, 42t411-419. Toccanier-Pelte, M., Skalli, O . , Kapanci, Y., and Gabbiani, G. (1987) Characterization of stromal cells with myoid features in lymph nodes and spleen in normal and pathological conditions. Am. J. Pathol., 129t109-118. Towbin, H., Staehelin, T., and Gordon, J. (1979) Electrophoretic transfer of protein from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. USA, 76t43504354. Viebahn, C., Lane, E., and Ramaekers, C. (1988) Keratin and vimentin expression in early organogenesis of the rabbit embryo. Cell Tissue Res., 253t553-562. Voyta, J., Via, D., Butterfield, C., and Zetter, B. (1984) Identification and isolation of endothelial cells based on their increased uptake of acetylated-low density lipoprotein. J. Cell Biol., 99t2034-2040. Weibel, E. (1980) Design and structure of the human lung. In: Pulmonary Diseases and Disorders. E. Fishman, ed. McGraw-Hill, New York, pp. 224-274. Weinberg, K., Douglas, W., MacNamee, D., Lanzillo, J., and Fanburg, B. (1982) Angiotensin I converting enzyme localization on cultured fibroblasts by immunofluorescence. In Vitro, 18:400-406. Woodcock-Mitchell, J., Burkhardt, A,, Mitchell, J., Rannels, S., Rannels, D., Chiu, J., and Low, R. (1986) Keratin species in type I1 pneumocytes in culture and during lung injury. Am. Rev. Respir. Dis. 134566-571. Wu, Y., Parker, L., Binder, N., Beckett, M., Sinard, J., Griffiths, T., and Rheinwald, J. (1982) The mesothelial keratins: A new family of cytoskeletal proteins identified in cultured mesothelial cells and nonkeratinizing epithelia. Cell, 3It693-703.
Expression of Simple Epithelial Cytokeratins in Bovine Pulmonary Microvascular Endothelial Cells WAYNE F. PATTON, MIN UNC YOON, J.STEVEN ALEXANDER, NANCY CHUNC-WELCH, HERBERT B. HECHTMAN, AND D A V I D SHEPRO" Biological Science Center, Boston University, Boston, Massachusetts 022 75 (W.F.P., M.U.Y., I.S.A., N.C.-W., D.S.); and Harvard Medical School, Boston, Massachusetts 02 I60 (H.B.H.)
Polypeptides of bovine aortic, pulmonary artery, and pulmonary microvascular endothelial cells, as well as vascular smooth muscle cells and retinal pericytes were evaluated by two-dimensional gel electrophoresis. The principal cytoskeleta1 proteins in all of these cell types were actin, vimentin, tropomyosin, and tubulin. Cultured pulmonary microvascular endothelial cells also expressed 12 unique polypeptides including a 41 kd acidic type I and two isoforms of a 52 kd basic type I I simple epithelial cytokeratin. Pulmonary microvascular endothelial cell expression of the simple epithelial cytokeratins was maintained in culture in the presence or absence of retinal-derived growth factor, and regardless of whether cells were cultured on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic. Cytokeratin expression was maintained through at least 50 population doubling in culture. The expression of cytokeratins was found to be regulated by cell density. Pulmonary microvascular endothelial cells seeded at 2.5 x lo5 cells/cm' (confluentseeding)expressed 3.5 times more cytokeratins than cells seeded at 1.25 x l o 4 cells/cm' (sparseseeding). Vimentin expression was not altered by cell density. By indirect immunofluorescence microscopy it was determined that the cytokeratins were distributed cytoplasmically at subconfluent cell densities but that cytokeratin 19 sometimes localized at regions of cell-cell contact after cells reached confluence. Vimentin had a cytoplasmic distribution regardless of cell density. These results suggest that pulmonary microvascular endothelial cells have a distinctive cytoskeleton that may provide them with functionally unique properties when compared with endothelial cells derived from the macrovasculature. In conjunction with conventional endothelial cell markers, the presence of simple epithelial cytokeratins may be an important biochemical criterion for identifying pulmonary microvascular endothelial cells. Embryologically, epithelial cells are derived from ectoderm, endoderm, and mesoderm. Endothelial cells are mesodermally derived simple squamous epithelial cells that line the heart, blood vascular, and lymphatic channels. Until recently, i t was thought that endothelial cells only contained the intermediate filament protein, vimentin. While in general this appears to be true, a n increasing number of exceptions have been found in microvascular endothelial cells. Desmin was demonstrated immunocytochemically to be the sole intermediate filament protein in chicken capillary endothelial cells including those of the renal cortex peritubular capillary and the hepatic sinusoid (Fujimoto and Singer, 1986). The chicken exocrine pancreas capillary was found to contain both vimentin and desmin (Fujimot0 and Singer, 1986). It has also been determined by immunocytochemistry that vimentin and desmin are coexpressed in the high endothelium venules of lymph nodes in rat but not in man (Toccanier-Pelte et al., 1987). 6) 1990 WILEY-LISS, INC.
The characteristic intermediate filament components of many epithelial cells are the cytokeratins, a multigene family of 19 or more polypeptides (Moll et al., 1982). They range in molecular weight from 40 t o 65 kd and in isoelectric point from 5.0 to 8.5. Two subfamilies of cytokeratins are distinguishable on the basis of the isoelectric points of the polypeptides: type I (acidic) cytokeratins and type I1 (basic) cytokeratins (Franke et al., 1981; Lazarides, 1982; Steinert et a]., 1985). Unlike vimentin and desmin, cytokeratins are obligatory heteropolymers requiring at least one member of each subfamily to form the tetramer subunit necessary for construction of the intermediate filament (Hatzfeld and Werner, 1985). The ratio of type I to type I1 cytokeratin in the intermediate filament is 1:l (Hatzfeld and Werner, 1985). Cytokeratins are also Received October 21, 1988; accepted December 8, 1989.
*To whom reprint requests/correspondence should be addressed.
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS
classified according to their cell type of origin. Cytokeratins 1-6 and 9-17 are commonly found in stratified epithelial cells while cytokeratins 7, 8, 18, and 19 are commonly associated with simple, nonkeratinizing epithelia (Bosch et al., 1988). Cytokeratins have been detected immunologically and electrophoretically in 4-8-cell-stage embryos, morulas, and blastocysts (reviewed in Lehtonen, 1987). In the trophectoderm cells of the blastocyst they are associated with desmosomes (Lehtonen, 1987). Coexpression of cytokeratins and vimentin has been demonstrated in a variety of embryonic tissues including the splanchnopleuric mesenchyme of the rabbit embryo (Viebahn et al., 1988). The simple epithelial keratins, cytokeratins 8 and 18, were found by immunocytochemistry with vimentin in synovial and submucosal microvascular endothelial cells of man (Jahn et al., 1987). In lower vertebrates, like the African clawed toad, all endothelial cells appear to contain both vimentin and simple epithelial keratins (Jahn et al., 1987). This widespread distribution of the simple epithelial cytokeratins may indicate that they should be considered as markers of differentiated cellular function rather than as germ layer derivation markers. In this study we demonstrate that bovine pulmonary microvascular endothelial cells, unlike any of the other vascular wall cell types examined, express simple epithelial cytokeratins. Since cytokeratins have different chemical and physical properties than other intermediate filaments like vimentin and desmin, finding this protein in the pulmonary microvascular endothelial cell may be important to this cell’s performance of differentiated functions, particularly with respect to maintaining a tight interendothelial barrier.
MATERIALS AND METHODS Isolation and culturing of cells Bovine pulmonary arteries, aortas, lungs, and retinas were obtained from a local slaughterhouse. Retinal pericytes were derived from isolated bovine retinal microvessels a s described by Gitlin and D’Amore (1983). Vascular smooth muscle cells were obtained from explant outgrowths of bovine aortas according to the method of Ross (1971). Rat pulmonary microvascular endothelial cells were isolated by a previously described method (Davies et al., 1987). A cell line of r a t pulmonary microvascular endothelial cells, isolated by retrograde perfusion of microcarriers, was obtained from Dr. U. Ryan. The macrovascular endothelial cells were obtained by a scraping technique a s previously described (Chung-Welch et al., 1988; Shepro et al., 1974). Pulmonary microvessel endothelial cells were obtained from peripheral lung sections and selected for over other cell types such a s pericytes and fibroblasts by differential plating as described previously (ChungWelch e t al., 1988). All cells were cultured in DulbecCO’SModified Eagle’s medium (DME) supplemented with 10% fetal bovine serum, 0.1% penicillin, 0.1% streptomycin, 0.1% amphotericin, and 0.1% glutamine. Primary cultures were passaged a t a 1:4 split ratio and cultured in 35 mm tissue culture dishes coated with 1.5%gelatin. For the studies evaluating the effects of different extracellular matrices on keratin expression, collagen I, collagen IV, fibronectin, tissue culture plas-
141
tic, laminin, or basement membrane proteins (Matrigel) were used according to manufacturer’s suggestions (Collaborative Research, Inc., Waltham, MA, USA).
Characterization of cell types Bovine pulmonary artery endothelial cells, zortic endothelial cells, and pulmonary microvascular endothelial cells expressed the typical endothelial cell “cobblestone” morphology and stringent contact inhibition and were found to stain positively with f luorescently labeled acetylated low-density lipoprotein, factor VIII, and angiotensin-converting enzyme (Chung-Welch et al., 1988; Weinberg et al., 1982; Voyta et al., 1984). We have also previously demonstrated SQ20881-inhibitable angiotensin-converting enzyme activity in all three of the endothelial cell types using [3H] benzoyl Phe-Ala-Pro as substrate (Chung-Welch et al., 1988). SQ20881 is a n angiotensin-converting enzyme antagonist (Peninsula Laboratories, Inc., San Pedro, CAI. In addition, the cells expressed the expected prostanoid patterns, with PGI, predominant in the macrovascular endothelial cells and either PGI, or PGE, predominant in the microvascular endothelial cells, dependent upon whether or not growth factor was included in the culture media (Chung-Welch et al., 1988). By visualization of the actin cytoskeleton with rhodamine-phalloidin i t was determined that the cells were responsive to vasoactive amines such as histamine and vasoactive peptides such a s angiotensin I1 and bradykinin. Smooth muscle cells were identified by their “hill and valley” morphology, lack of contact inhibition, and lack of staining with fluorescently labeled acetylated lowdensity lipoprotein. Retinal pericytes also did not stain with fluorescently labeled acetylated low-density lipoprotein, but did not form a “hill and valley” morphology and proliferated very slowly in culture. Metabolic labeling of polypeptides For the detection of total cellular protein, the normal 10% fetal calf serum-DMEM was replaced by 90% methionine-free FCS-DMEM, 10% normal FCS-DMEM supplemented with 0.1 mCilml of Tran”5S-label (I.C.N., Irvine, CA) just prior to endothelial cells reaching confluence. Cells were then incubated for 24-48 h in the labeling media. T ~ - a n ~ ~ S - l aisb used e l a s a [35Sl-methionine substitute for metabolic labeling. It contains 709k [35Sl-methionine and 20% [35Sl-cysteine. Smooth muscle cells were also labeled for 24-48 h whereas retinal pericytes due to their slow proliferation rate were generally labeled for 72 h. Cell lysate and cytoskeleton isolation Triton-insoluble cytoskeletons were prepared by a modification of the method of Gilbert and Fulton (1985). All buffers were maintained a t 37°C during the cytoskeleton isolation. Endothelial cell monolayers in 35 mm dishes were washed twice in DMEM and then extracted for 1 min in 10 mM Hepes, 100 mM KCl, 2.5 mM MgCI,, 1 mM CaC12, 300 mM Sorbitol, 1.0 mM phenylmethylsulfonyl fluoride (PMSF), 1.0%Triton X100, adjusted to pH 7.4 (extraction buffer). The extraction buffer was carefully aspirated and the insoluble structures remaining associated with the dish were incubated for a n additional minute with fresh extraction buffer. After removing the extraction buffer by aspira-
142
PATTON ET AL.
tion, a final wash was performed using 2 ml of 10 mM Tris-C1, 1.0% Triton X-100, 1.0 mM PMSF, adjusted to pH 7.4. The Triton-insoluble cytoskeletons were solubilized by addition of 60 p1 of 95°C 2.5% sodium dodecylsulfate (SDS), 20 mM Tris-C1, 1.0 mM PMSF, 2.5% B-mercaptoethanol, adjusted to pH 8.0. By using the point of a 200 pl pipet tip, the monolayer was vigorously scraped until a n amorphous, transparent, sticky layer was released from the bottom of the dish corresponding to the cytoskeleton. In order to obtain cell lysates, monolayers were washed twice in DMEM and then solubilized by addition of 100 pl of 95°C2.5% SDS, 20 mM Tris-C1, 1.0 mM PMSF, 2.5% B-mercaptoethanol, adjusted to pH 8.0. Monolayers were then scraped in a similar manner to the Triton-insoluble cytoskeletons.
molecular weight standards (Sigma Chemical Co., St. Louis, MO). Though molecular weights and isoelectric points were determined by using external standards, the values obtained for actin and vimentin were found to be within 3% of the literature values.
Immunofluorescence microscopy For immunofluorescence studies, endothelial cells were grown to confluency on 12 mm round glass coverslips that had previously been coated with 1.5% gelatin. Cells were fixed in 3.7% phosphate-buffered formaldehyde (pH 7.4) for 15 min. The cells on coverslips were permeabilized in extraction buffer (0.5 M KC1, 1% Triton X-100, 10 mM MgCl,, 1 mg/ml Tris HC1, 17 pg/ml TSF, 0.25 mg/ml DNAse 1 in phosphate-buffered saline (PBS)) for 10 min. This solution was removed and the cells were washed three times with PBS. The Protein electroblotting and visualization of the cells were stained with monoclonal anti-cytokeratin cytokeratins and vimentin 4.62, monoclonal anti-cytokeratin 8.12 or monoclonal The cytokeratins and vimentin were detected immu- anti-vimentin 13.2 antibody. FITC-conjugated rabbit nologically by transferring proteins to nitrocellulose anti-mouse IgG was added to visualize the intermedisheets by the method of Towbin et al. (1979). Prior to ate filaments by immunofluorescence microscopy. The electroblotting, the gels were allowed to pre-equil- coverslips were washed five times in PBS, mounted in ibrate at room temperature in electroblot buffer (50 PBS/glycerol (l:l), and sealed. Cells on coverslips were mM Tris-base, 95 mM glycine, 0.005% SDS) for 30-45 illuminated for fluorescence by using a Zeiss universal min. The transfer to nitrocellulose paper was per- microscope with a n Olympus OM-25 camera attachformed for 3 h a t 250 mA constant current and 40-50 ment. volts. The blots were incubated overnight with polyRESULTS clonal antibody produced by using bovine hooves as immunogen (a general anti-keratin antibody), monoComparison of polypeptide patterns of clonal anti-cytokeratin antibody (clone no. K4.62, spe- pulmonary microvascular endothelial cells with cific for cytokeratin 19), monoclonal anti-cytokeratin other cells of the vascular wall antibody (clone K8.12, reacts with a broad range of The distribution of polypeptides in pulmonary arkeratins including cytokeratin 81, or monoclonal antivimentin antibody (clone no. VIM-13.2, specific for vi- tery, aortic, and pulmonary microvascular endothelial mentin). The blots were then incubated with alkaline- cells, a s well a s vascular smooth muscle cells and retphosphatase-conjugated rabbit anti-mouse IgG or inal pericytes was examined by two-dimensional gel rabbit anti-guinea pig IgG diluted 1:1,000, and visu- electrophoresis. All cell types examined possessed the alized by using a nitroblue tetrazolium chloride/ cytoskeletal proteins actin, vimentin, tubulin, and 5-bromo-4-chloro-3-indoylphosphate p-toluidine chro- tropomyosin a s determined by their characteristic promogen system. All antibodies and chromogens were tein mobilities reported by others (Savion et al., 1982; Blose, 1984). Actin was also identified by its comigrapurchased from Sigma Chemical Co., St. Louis, Mo. tion with rabbit skeletal muscle actin while vimentin Two-dimensional SDS polyacrylamide gel was identified by using monoclonal anti-vimentin 13.2 electrophoresis antibody. Though each cell type was observed to have Two-dimensional gel electrophoresis was performed polypeptides not possessed by the other cell types, the according to the method of O'Farrell (1975). To reduce most striking polypeptide pattern differences were obstreaking in the isoelectric focusing dimension, the tained from pulmonary microvascular endothelial samples were prepared as described by Ames and Nai- cells. Approximately 340 polypeptides were detected by kaido (1976). The second dimension was run on 10% quantitative two-dimensional gel electrophoresis of polyacrylamide slab gels and for Tran3%-1abel studies, whole cell lysates from pulmonary microvascular ensubsequently incubated for 30 min in two changes of dimethylsulfoxide, followed by a 3 h incubation in 20% 2,5-diphenyloxazole (PPO) in dimethylsulfoxide, a 1 h incubation in water, and then a 2 h incubation in 2.5% Fig. 1. Two dimensional gel electrophoretic comparison of various glycerol prior to air drying between dialysis sheets cells of the vascular wall after metabolic labeling with Tran"5S-label. (Dunbar, 1987). Radiolabeled gels were exposed to A Retinal pericytes. B: Vascular smooth muscle cells. C: Pu1mona:ry Kodak X-OMAT film by manufacturer's instructions. artery endothelial cells. D Aortic endothelial cells. E: Pulmonary endothelial cells. F: Consensus map of polypeptides Gels were analyzed by using a microcomputer-based microvascular identified in pulmonary microvascular endothelial cells. Proteins in videodensitometer as previously described (Mariash et black were found to be unique to pulmonary microvascular endotheal., 1982; Patton e t al., 1989). Isoelectric points were lial cells. The other proteins shown in the map are expressed in macdetermined by measuring the pH of 1 cm segments of rovascular endothelial cells, smooth muscle cells, and pericytes ;as well as the pulmonary microvessel endothelial cells. Many are exblank gels run in parallel with the sample while mo- pressed in different amounts between the cell types. Two of the unique lecular weights were determined from one-dimensional polypeptides corresponded to cytokeratins 8 and 19. Vimentin and the electrophoretic separations of commercially available three principal isoforms of actin (a$,?)are also identified in the map.
Fig. 1.
144
PATTON ET AL.
dothelial cells. The average abundance of each polypeptide resolved on the gels was found to be 0.29%. Twelve major polypeptides, each representing 0.21 to 1.7% of the total cellular protein, were found to be expressed exclusively by the pulmonary microvascular endothelial cells. To identify further the polypeptides expressed uniquely in the pulmonary microvascular endothelial cells, Triton X-100-resistant cytoskeletons were isolated and evaluated. Three of the unique polypeptides observed in the cell lysate patterns were enriched substantially in the cytoskeleton preparation. The resistance of these polypeptides to extraction with Triton X-100 as well as their high abundance in the cell are properties consistent with them being intermediate filament proteins (Moll et al., 1982). The molecular weights and isoelectric points of the three polypeptides were M, = 41 kd, PI = 5.1; M, = 52 kd, PI = 5.65; and M, = 52 kd, PI = 5.9. The two polypeptides with the same molecular weight are believed to represent different isoforms of the same polypeptide. The molecular weight and isoelectric point coordinates of the pulmonary microvascular endothelial cell proteins corresponded t o the coordinates of known simple epithelial cytokeratins (Wu et al., 1982). These polypeptides correspond to cytokeratins 19 and 8 when one uses the nomenclature of Moll et al. (1982). Figure 1 shows a representative comparison made between whole cell lysate polypeptide patterns of all the cell types examined as well a s the positions of actin, vimentin, and the cytokeratins. The relative abundances of the cytokeratins, actin and vimentin were determined in pulmonary microvascular endothelial cell lysates and cytoskeletons by quantitative videodensitometry. In whole cell lysates actin represented 4.5%,vimentin represented 1.2%, cytokeratin 19 represented 1.14%, and the two isoforms of cytokeratin 8 together represented 1.7% of the total amount of protein visualized on the two-dimensional gels. The experimentally determined ratio of cytokeratin 19 to cytokeratin 8 was thus found to be 1:1.5. A comparison of the nucleotide sequences of bovine keratins 8 and 19 obtained from the National Gene Bank revealed that cytokeratin 8 contains 15 methionine residues while cytokeratin 19 contains 10. Correcting for the greater number of methionine residues in cytokeratin 8 yields a cytokeratin 8 to 19 ratio of 1:1, consistent with the known requirements for intermediate filament assembly discussed in the Introduction. In l’riton-insoluble cytoskeletons, actin represented 10.9%,vimentin represented 4.3%, cytokeratin 19 represented 3.5%, and the two isoforms of cytokeratin 8 represented 5.3% of the total amount of protein visualized on the two-dimensional gels. For comparison, i t has previously been demonstrated by two-dimensional gel electrophoresis that actin represents 3.3%, vimentin 0.55%, and each of the five keratin species represent from 0.07 to 0.63% of the total amount of visualized protein in HeLa cell lysates (Bravo and Celis, 1984). Vimentin was found to be present in pulmonary microvascular endothelial cells in only 33%of the amount found in pulmonary artery endothelial cells. The total amount of intermediate filament protein in pulmonary microvascular endothelial cells was slightly greater than pulmonary artery endothelial cells.
Immunological characterization of the simple epithelial cytokeratins and vimentin Immunological analysis with polyclonal antibody raised against epidermal keratins, monoclonal antibody reactive to a broad range of keratins, and monoclonal antibody reactive to cytokeratin 19 confirmed the identification of the polypeptides as simple epithelial cytokeratins. The identity of vimentin was also confirmed by using monoclonal antibody raised against vimentin. Western blots probed with the anti-keratin polyclonal antibody showed reactivity with a number of polypeptides in both aortic endothelial cells and pulmonary microvascular endothelial cells, including some soluble, noncytoskeletal proteins, but reactivity with the 41 kd and 52 kd polypeptides (cytokeratins 19 and 8 ) was restricted to the pulmonary microvascular endothelial cells (data not shown). A monoclonal antibody reactive to a broad range of keratins (K8.12) was found to display low crossreactivity with soluble proteins. As shown in Figure 2, monoclonal antibody K8.12 selectively reacted with cytokeratin 8 and 19 i~n pulmonary microvascular endothelial cells. The monoclonal antibody to cytokeratin 19 (K4.62) primarily reacted with the 41 kd protein in pulmonary microvascular endothelial cells while vimentin was detected in both endothelial cell types. Two-dimensional gels transferred to nitrocellulose sheets and probed with monoclonal antibody K8.12 also demonstrated the presence of cytokeratins 8 and 19 in pulmonary microvascular endothelial cells (see Fig. 3). As shown in Figure 4, immunofluorescence microscopy of monolayers of cells with monoclonal anti-cytokeratin 19, monoclonal anti-cytokeratin (K8.12), and monclonal antivimentin (VIM-13.12) antibody demonstrated that both keratin and vimentin intermediate filaments existed in the pulmonary microvascular endothelial cell cytoplasm as distinct filament networks. At subconfluence the cytokeratins showed a cytoplasmic distribution, but after the cells were cultured for extended periods of time a t confluency (1-3 weeks), cytokeratin 19 was often found to be localized a t the periphery of the cell and appeared to be associated with the cell-cell junctions. Vimentin had a more diffuse distribution in the cytoplasm regardless of the state of confluenc,y. Aortic endothelial cells had a similar distribution of vimentin but the cytokeratins were not detected in the cells. Expression of the simple epithelial cytokeratins was observed to be maintained in culture in the presence or absence of retinal-derived growth factor, and regardless of whether cells were cultured on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic through a t least 50 population doublings. The levels of keratin expression were found to be largely independent of the substrate pulmonary microvascular endothelial cells were exposed to. Cells seeded on gelatin, fibronectin, collagen I, collagen IV, laminin, basement membrane proteins, or plastic a t 1:4 split ratios (6.25 x lo4 cells/cm2) and allowed to reach confluency differed in the amount of expressed keratin by less than 10-15% relative to the other cytoskeletal proteins (n = 3). The levels of keratins expressed in pulmonary microvascular endothelial cells were found to be profoundly influenced by cell
CYTOKERATTNS IN PULMONARY ENDOTHELIAL CELLS
145
Fig. 2. Immunological identification of the cytokeratins in pulmonary microvascular endothelial cells. Whole cell lysates of pulmonary microvascular and aortic endothelial cells were resolved by one-dimensional gel electrophoresis and either stained with Coomassie blue dye or electrophoretically transferred to nitrocellulose for subsequent probing with antibodies as described in the Methods section. A Coomassie blue dye patterns of microvascular and aortic endothelial cell proteins. B: Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal antibody K8.12 (reactive to a variety of acidic and basic keratins). Cytokeratins 8 and 19 were labeled only in the microvascular endothelial cells. C.
Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal anti-cytokeratin antibody (clone no. K4.62, specific for cytokeratin 19). A minor second band that is less than 10% of the keratin 19 band is also labeled with the K4.62 antibody. On 2 D gels this second band has a very acidic isoelectric point (4.6) but remains unidentified. D Electroblotted lanes of microvascular and aortic endothelial cell proteins immunostained with monoclonal anti-vimentin antibody (clone no. VIM-13.2). The positions of vimentin (V), cytokeratin 8 (C 81, actin ( A ) ,and cytokeratin 19 (C 19) are indicated.
density. Cells seeded at 2.5 x lo5 cells/cm2 (confluent seeding) expressed 3.5 times more cytokeratins 8 and 19 than cells seeded a t 1.25 x lo4 cells/cm2 (sparse seeding) (n = 3). As the cells became progressively more confluent, the amount of keratin in their cytoplasms steadily increased.
cardial, and peritoneal cavities as well as the outer surfaces of the lungs, heart, and viscera. Human mesothelial cells have been shown to express cytokeratins 7, 8, 18, and 19 (Wu et al., 1982; Connell and Rheinwald, 1983; Kim et al., 1987). Thus far, cytokeratin 7 has not been shown to be exmessed in bovine cells (O'Guin et al., 1987). Mesothelial cells and endothelial cells are difficult to distinguish from one another by histological criteria alone (Rheinwald et al., 1987; Satoh and Prescott, 1987). Endothelial cells present a nonthrombogenic, nonadhesive surface to the blood whereas mesothelial cells perform the same function with respect to the serous fluids (Rheinwald e t al., 1987; Satoh and Prescott, 1987). Our pulmonary microvascular endothelial cells as well as mesothelial cells have numerous microvilli projections on the apical plasma membrane surface (Chung-Welch et al., 1988; Andrews and Porter, 1973; Satoh and Prescott, 1987). Both bovine pericardial mesothelial cells and bovine pulmonary microvascular endothelial cells stain positively for Factor VIII-related antigen (Satoh and Prescott, 1987; Chung-Welch et al., 1988). We demonstrated that unstimulated cultures of pulmonary microvascular endothelial cells synthesized 0.86 ng PGI,/mg protein while 1.0 pM bradykinin and
DISCUSSION In this study, cultured bovine pulmonary microvessel endothelial cells were found to express cytokeratins 8 and 19 a s well a s the intermediate filament protein vimentin. A number of epithelial cell types including hepatocytes, colonic mucosa, and urothelium contain . for a both vimentin and keratin (see Moll et ~ 1 .1982, more extensive listing). This pattern of intermediate filament protein expression, however, is exceedingly rare among cells derived from the mesodermal germ layer and thus pulmonary microvascular endothelial cells belong to a novel class of mesodermally derived cells expressing simple epithelial-type cytokeratins as well as vimentin (Wu et al., 1982). Mesothelial cells are the only other cells derived from the mesoderm known to synthesize vimentin in conjunction with simple epithelial keratins. Mesothelial cells are the flat cells contained in the single layer lining the pleural, peri-
146
PATTON ET AL
Fig. 3. Detection of cytokeratins 8 and 19 by two-dimensional gel electrophoresis followed by Western blotting and incubation with monoclonal antibody K8.12. A Two-dimensional gel of cytoskeletal proteins detected after labeling with "ran%-label. B Two-dimen-
sional Western blot after reaction with monoclonal antibody K8.12. The positions of vimentin (V), cytokeratin 8 (C €3, actin (A), and cytokeratin 19 (C 19)are indicated.
1.9 pM of the calcium ionophore A23187 caused 2.14 and 2.66 ng PGI,/mg protein to be synthesized in the cultures respectively (Chung-Welch et al., 1988). Others have demonstrated that cultured bovine pericardial mesothelial cells synthesized 2.1 ng PGI,/mg protein whereas 10 p M of bradykinin and 5 p M of the calcium ionophore A23187 caused 18.3 and 94.6 ng PGI,/mg
protein to be synthesized in the cultures respectively (Satoh and Prescott, 1987). In humans, mesothelial cells also share with endothelial cells the expression of a prominent secreted glycoprotein (M, 46 kD, PI 7.11, commonly referred to as mesosecrin (Rheinwald et al., 1987). It should be mentioned a t this point that the pulmonary microvascular
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS
Fig. 4. Immunofluorescence localization of cytokeratin and vimentin in subconfluent pulmonary microvascular and aortic endothelial cells. A Pulmonary microvascular endothelial cells labeled with antivimentin antibody (clone no. VIM-13.2). B: Aortic endothelial cells labeled with anti-vimentin antibody (clone no. VIM-13.2). C: Pulmonary microvascular endothelial cells labeled with anti-cytokeratin antibody (clone no. K4.62,specific for cytokeratin 19). D Aortic endothelial cells labeled with anti-cytokeratin antibody (clone no.
147
K4.62). Only very weak nonspecific background staining by the rhodamine-conjugated second antibody was detectable. E: Pulmonary microvascular endothelial cells labeled with anti-cytokeratin antibody (clone no. K8.12,specific for a broad spectrum of cytokeratins including cytokeratin 8 and 18). F: Aortic endothelial cells labeled with anti-cytokeratin antibody (clone no. K8.12). Only very weak nonspecific background staining by the rhodamine conjugated second antibody was detectable. Bars, 1 p,m.
148
PATTON ET AL
endothelial cells described in this study were isolated after physical removal of the lung pleura, a potential source of mesothelial cells (Chung-Welch et al., 1988). The cell cultures obtained from the dissected peripheral lung by filtering with Nitex screens and differential plating were examined by using a battery of markers to confirm their identity as endothelial cells (Chung-Welchet al., 1988).We have examined rat pulmonary microvascular endothelial cells isolated by the filtering method of Davies et al. (1987) and by the retrograde perfusion method of Ryan et al. (1982). Both of these cell lines contained an abundance of cytokeratins 8 and 18 as well as some cytokeratin 19 as determined by two-dimensional gel electrophoresis (Dodge et al., submitted). We could not detect cytokeratin 18 in bovine pulmonary microvascular endothelial cells by twodimensional gel electrophoresis of cytoskeletal proteins or by immunoreactivity with the general anti-keratin antibody. Using high-resolution two-dimensional gels we have been able to detect cytokeratin 18 in bovine pericardial mesothelial cells. The pulmonary microvascular endothelial cells can also be distinguished from mesothelial cells on the basis of their ability to form tube-like structures on basement membrane proteins or collagen I (Chung-Welch et al., 1989). The bronchial tree is a meticulously designed system of branching airways containing an abundance of epithelial cell types (Weibel, 1980). Few studies concerning the keratin patterns of these cells have been conducted due to the difficulty in obtaining pure preparations, free of contaminating cell types. Moll et al. (1982) found cytokeratins 6-8 and 17-19 in human undifferentiated carcinoma of the lung and they observed cytokeratins 5-8, 13-15, and 17-19 in human tracheal epithelium. This is consistent with the cells originating from stratified epithelia. Blobel et al. (1984)found the simple epithelial cytokeratins 7,8,18, and 19 in the alveolar regions of the human lung by microdissection techniques and immunofluorescence microscopy. They were unable to assign these species to individual cell types in this region of the lung, however. Woodcock-Mitchell et al. (1986) examined pure cultures of Type I1 alveolar pneumocytes as well as paraffin-embedded sections of normal and injured rat lung for cytokeratin species. Type I1 alveolar pneumocytes contained cytokeratins 7, 8, and 18 but no cytokeratin 19. None of the lung epithelial cell types described expressed more than trace amounts of the intermediate filament protein, vimentin. The basic structural property of lung parenchyma is the reduction of the tissue mass to a thin sheet that provides the air and blood an extensive access to one another (Weibel, 1980). The mechanical problems inherent in suspending the delicate alveolar septa while exposing the capillary to air over a large surface require a highly organized connective tissue lattice. Ventilatory movements of the thorax and diaphragm are transmitted all along the bronchial tree through the peripheral fiber lattice system to the alveoli themselves (Weibel, 1980). The capillary endothelium as a minimal cell layer constituting a barrier between blood and surrounding tissue may require an epithelial-like intermediate filament network to compensate for the unusual mechanical burden it is subjected to by virtue of its anatomical location. Cytokeratins may function
as the intracytoplasmic portion of desmosomes, providing anchorage points between individual cells and distributing mechanical stress applied to one cell over the entire tissue. In situ analysis of human and bovine lung using monoclonal antibody specific to cytokeratin 19 demonstrates that this intermediate filament protein is associated with endothelial junctions in some microvascular beds while being cytoplasmically distributed in others (Mineau-Hanske et al., submitted). In this study we demonstrate that cultured bovine pulmonary microvascular endothelial cells, unlike any of the other vascular wall cell types examined, express two simple epithelial cytokeratins: an acidic type I keratin (M, = 41 kd, PI = 5.1) and a basic type I1 keratin (M, = 52 kd, PI = 5.7-5.9). Since cytokeratins have different chemical and physical properties than other intermediate filaments like vimentin and desmin, finding this protein in the pulmonary microvascular endothelial cell may be important to this cell’s performance of differentiated functions.
ACKNOWLEDGMENTS The authors wish to acknowledge personnel from our laboratory: Andrea Dodge for the retinal pericyte cultures; Jeanne Choi for assistance with fluorescence microscopy; and G.P. Ameia Yen-Patton, Dr. Rochelle Mineau-Hanske, and Dr. Nicole Morel for useful discussions. This investigation was performed at Boston University, Biology Department, Boston, MA 02215, and was supported in part by USPHS grants HBL16714,33104, and GM24891.
LITERATURE CITED Ames, G., and Nikaido, K. (1976)Two-dimensional gel electrophoresis of membrane proteins. Biochemistry, 15:616-623. Andrews, P., and Porter, K. (1973) The ultrastructural morphology and possible functional significance of mesothelial microvilli. Anat. Rec., 177:409-426. Blobel, G., Moll, R., Franke, W., Vogt-Moykopf, I. (1984) Cytokeratins in normal lung and lung carcimonas. Virchows. Arch. l Cell Pathol. J 45:407-429. Blose, S. (1984) The endothelial cytoskeleton. In: Biology of Endothelial Cells. E.M. Jaffee, ed. Martinus, Boston, Ch. 14, pp. 141-154. Bosch, F., Leube, R., Achtstatter, T., Moll, R., and Franke, W. (1988) Expression of simple epithelial type cytokeratins in stratified epithelia as detected by immunolocalization and hybridization in situ. J . Cell Biol., 106t1635-1648. Bravo, R., and Celis, J. (1984) Catalog of HeLa cell proteins. In: TwoDimensional Gel Electrophoresis of Proteins; Methods and Applications. J. Celis and R. Bravo, eds. Academic Press, New York, Ch. 14,pp. 445-476. Chung-Welch, N., Shepro, D., Dunham, B., and Hechtman, H. (1988) Prostacyclin and prostaglandin E, secretions by bovine pulmonary microvessel endothelial cells are altered by changes in culture conditions. J. Cell. Physiol., 135t224-234. Chung-Welch, N., Patton, W., Yen-Patton, G., Hechtman, H., and Shepro, D. (1989) Phenotypic comparison between mesothelial and microvascular endothelial cell lineages using conventional endothelial cell markers, cytoskeletal protein markers and in uitro assays of angiogenic potential. Differentiation 42:44-53. Connell, N., and Rheinwald, J. (1983) Regulation of the cytoskeleton in mesothelial cells: Reversible loss of keratin and increase in vimentin during rapid growth in culture. Cell, 34t245-253. Davies, P., Smith, B., Maddalo, F., Langleben, D., Tobias, D., Fujiwara, K., and Reid, L. (1987) Characterization of lung pericytes in culture including their growth inhibition by endothelial substrate. Microvasc. Res., 33t300-314. Dodge, A., Patton, W., Yoon, M., Hechtman, H., and Shepro, D. (1989) Organ and species specific differences in intermediate filament protein profiles of cultured microvascular endothelial cells. J. Cell Sci. (submitted).
CYTOKERATINS IN PULMONARY ENDOTHELIAL CELLS Dunbar, B. (1987) Two-Dimensional Electrophoresis and Immunological Techniques. Plenum Press, New York. Eichner, R., Sun, T., and Aebi, U. (1986) The role of keratin subfamilies and keratin pairs in the formation of human epidermal intermediate filaments. J. Cell Biol., 102r1767-1777. Franke, W., Schiller, D., Moll, R., Winter, S., Schmid, E., and Engelbrecht, I. (1981) Diversity of cytokeratins. Differentiation specific expression of cytokeratin polypeptides in epithelial cells and tissues. J . Mol. Biol., 153t933-959. Fujimoto, T., and Singer, S. (1986) Immunocytochemical studies of endothelial cells in vivo. I. The presence of desmin only, or of desmin plus vimentin, or vimentin only, in the endothelial cells of different capillaries of the adult chicken. J. Cell Biol., 103r27752786. Gilbert, M., and Fulton, A. (1985) The specificity and stability of the triton-extracted cytoskeletal framework of gerbil fibroma cells. J. Cell Sci., 73:335-415. Gitlin, J., and D’Amore, P. (1983) Culture of retinal capillary cells using selective growth media. Microvasc. Res., 26t74-80. Hatzfeld, M., and Franke, W. (1985) Pair formation and promiscuity of cytokeratins: Formation in vitro of heterotypic complexes and intermediate-sized filaments by homologous and heterologous recombinations of purified polypeptides. J. Cell Biol., 1OIt1826-1841. Jahn, L., Fouquet, B., Robe, K., and Franke, W. (1987) Cytokeratins in certain endothelial and smooth muscle cells of two taxonomically distant vertebrate species, Xenopus laevis and man. Differentiation, 36:234-254. Kim, K., Stellmach, V., Javors, J., and Fuchs, E. (1987) Regulation of human mesothelial cell differentiation: Opposing roles of retinoids and epidermal growth factor in the expression of intermediate filament proteins. J. Cell Biol., 105t3039-3051. Lazarides, E. (1982) Intermediate filaments: A chemically heterogeneous, developmentally regulated class of proteins. Annu. Rev. Biochem., 51t219-250. Lehtonen, E. (1987) Cytokeratins in oocytes and preimplantation embryos in the mouse. In: Current Topics in Developmental Biology, Volume 22: The Molecular and Developmental Biology of Keratins. R. Sawyer, ed. Academic Press, New York, Chapter 7, pp. 153-173. Mariash, C., Seelig, S., and Oppenheimer, S. (1982) A rapid, inexpensive, quantitative technique for the analysis of two-dimensional electrophoretograms. Anal. Biochem., 12It388-394. Mineau-Hanschke, R., Patton, W., Hechtman, H., and Shepro, D. (1989) Immunolocalization of cytokeratin-19 in bovine and human pulmonary microvessel endothelial cells in situ. In Vitro (submitted). Moll, R., Franke, W., and Schiller, D. (1982) The catalog of human cutokeratins: Patterns of expression in normal epithelial, tumors and cultured cells. Cell, 31:ll-24. O’Farrell, P. (1975) High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem., 25Ot4007-4021. O’Guin, M., Galvin, S., Schermer, A,, and Sun, T. (1987) Patterns of keratin expression define distinct pathways of epithelial development and differentiation. In: Current Topics in Developmental Biology, Volume 22: The Molecular and Developmental Biology of
149
Keratins. R. Sawyer, ed. Academic Press, New York, Chapter 5, pp. 97-125. Patton, W., Dhanak, M., and Jacobson, B. (1989) Identification of dictyostelium discoideum plasma membrane proteins by cell surface labeling and quantitative two-dimensional gel electrophoresis. Anal. Biochem., 179t37-49. Rheinwald, J., Jorgensen, J., Hahn, W., Terpstra, A,, O’Connell, T., and Plummer, K. (1987) Mesosecrin: A secreted glycoprotein produced in abundance by human mesothelial, endothelial and kidney epithelial cells in culture. J. Cell Biol., 104:263-275. Ross, R. (1971) The smooth muscle cell 11. Growth of smooth muscle in culture and formation of elastic fibers. J. Cell Biol., 50t172-186. Ryan, U., Clements, E., Habliston, D., and Ryan, J . (1982) Use of microcarriers to isolate and culture pulmonary microvascular endothelium. Tissue Cell, 16.577-691. Satoh, K., and Prescott, S. (1987) Culture of mesothelial cells from bovine pericardium and characterization of their arachidonate metabolism. Biochim. Biophys. Acta, 930:283-296. Savion, N., Vlodavsky, G., Greenburg, G . , and Gospodarowicz, D. (1982) Synthesis and distribution of cytoskeletal elements in endothelial cells as a function of cell growth and organization. J. Cell. Physiol., 110t129-141. Shepro, D., Rosenthal, M., Batbouta, L., Roblee, L., and Belamarich, F. (1974) The cultivation of aortic endothelium. Anat. Rec., 78:523. Steinert, P., Steven, A., and Roop, D. (1985) The molecular biology of intermediate filaments. Cell, 42t411-419. Toccanier-Pelte, M., Skalli, O . , Kapanci, Y., and Gabbiani, G. (1987) Characterization of stromal cells with myoid features in lymph nodes and spleen in normal and pathological conditions. Am. J. Pathol., 129t109-118. Towbin, H., Staehelin, T., and Gordon, J. (1979) Electrophoretic transfer of protein from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc. Natl. Acad. Sci. USA, 76t43504354. Viebahn, C., Lane, E., and Ramaekers, C. (1988) Keratin and vimentin expression in early organogenesis of the rabbit embryo. Cell Tissue Res., 253t553-562. Voyta, J., Via, D., Butterfield, C., and Zetter, B. (1984) Identification and isolation of endothelial cells based on their increased uptake of acetylated-low density lipoprotein. J. Cell Biol., 99t2034-2040. Weibel, E. (1980) Design and structure of the human lung. In: Pulmonary Diseases and Disorders. E. Fishman, ed. McGraw-Hill, New York, pp. 224-274. Weinberg, K., Douglas, W., MacNamee, D., Lanzillo, J., and Fanburg, B. (1982) Angiotensin I converting enzyme localization on cultured fibroblasts by immunofluorescence. In Vitro, 18:400-406. Woodcock-Mitchell, J., Burkhardt, A,, Mitchell, J., Rannels, S., Rannels, D., Chiu, J., and Low, R. (1986) Keratin species in type I1 pneumocytes in culture and during lung injury. Am. Rev. Respir. Dis. 134566-571. Wu, Y., Parker, L., Binder, N., Beckett, M., Sinard, J., Griffiths, T., and Rheinwald, J. (1982) The mesothelial keratins: A new family of cytoskeletal proteins identified in cultured mesothelial cells and nonkeratinizing epithelia. Cell, 3It693-703.
