In Vitro Cell. Dev. Biol. 28A:711 - 715, November-December 1992 © 1992 Tissue Culture Association 0883-8364/92 $01.50+0.00

CULTURE AND CHARACTERIZATION OF PULMONARY MICROVASCULAR ENDOTHELIAL CELLS PETER J. DEL VECCHIO, ALMA SIFLINGER-BIRNBOIM, PAULA N. BELLONI, LISA A. HOLLERAN, HAZEL LUM, AND ASRAR B. MALIK Departments of Ophthalmology (A-77) (P. J. DV.) and Physiology and Cell Biology (P. J. DV., A. S-B., L. A. H., H. L., A. B. M.), The Albany Medical College of Union University, 47 New Scotland Avenue, Albany, New York 12208; and Department of Tumor Biology (P. N. B.), M.D. Anderson Cancer Center, Houston, Texas 77030 (Received 26 February 1992; accepted 27 March 1992)

SUMMARY Surface proteins were compared in endothelial cells (EC) obtained from bovine peripheral lung, pulmonary artery and vein, and dorsal aorta using sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis. Galactose-containing glycoproteins [molecular weight (M~) 1 6 0 - 2 2 0 and 40 kDa] binding to the Ricinus communis agglutinin (RCA) and peanut agglutinin (PNA) were selectively observed on pulmonary microvessel EC as compared to EC from pulmonary artery, pulmonary vein, and dorsal aorta. The unique RCA- and PNA-binding profiles of EC from the pulmonary artery and microvessels may be important in characterizing EC from different sites in the pulmonary circulation. The pulmonary microvessel EC monolayer was also 15-fold more restrictive to transendothelial flux of [14C]sucrose (Mr = 342 Da) than the pulmonary artery EC monolayer. In contrast, the microvessel EC were only six- and twofold more restrictive to the flux of larger tracer molecules, ovalbumin (Mr 43 kDa) and albumin (Mr = 69 kDa) than pulmonary artery EC. The greater restrictiveness of pulmonary microvessel EC monolayer indicates a major pbenotypic difference in the cultured pulmonary microvessel EC barrier function. Key words: pulmonary microvessel endothelial cells; cell identification; surface glyeoproteins; lectin binding. Company, Mahwah, NJ) and rinsed 3 times with Ca2+- and Mg~+-freephosphate buffered saline (cmf-PBS) containing Fungizone (Flow Laboratories, McLean, VA) (2.5 mg/ml). The pleura was dissected away and small pieces of tissue from the lung periphery were removed and placed in a 60-mm tissue culture dish (Coming Glass Company, Coming, NY) and rinsed 3 times with cmf-PBS containing Fungizone. The tissue was minced thoroughly and tile pieces were suspended in 10 ml solution containing 1000 U/ml collageuase (CLS2) in 5% bovine serum albumin (BSA) in a sterile 50-ml Erlenmeyer screw-top flask and stirred for 30 rain at 37 ° C. The material was then filtered through nylon bolting cloth (openings of 160/.tin) (Tetko Inc., Elmsford, NY). The filtrate was centrifuged at 10 ° C and 200 Xg for 5 min and the tissue pellet was resuspended in 25 ml of medium with serum, rinsed, and centrifuged 3 times. After the final eentrifugation, the pellet was resuspended in RPMI 1640 and 20% fetal bovine serum (FBS) (HyClone Laboratories, Inc., Logan, UT) containing 75 ttg/mt heparin (Sigma, St. Louis, MO) and 6.7 #g/ml retinal derived growth factor (RDGF) (complete medium). The tissue suspension was placed in fibronectin (FN)-coated 100-ram dishes (Coming). The culture medium contained a mixture of 3-parts complete medium and 1-part complete medium that was conditioned by placing over cultures of human umbilical vein EC for 2 days. The conditioned medium had been filtered through a 0.22-tim Millex-GVd filter unit (Millipore Products Division, Bedford, MA). Gentamicin (Whittaker Bioproducts, Walkersville, MD) (50 gg/ml) was included in the medium for the first three feedings. Selection of endothelial cell colonies. After several days incubation at 37 ° C in 5% COs, small, isolated colonies of EC were observed (Fig. 1 a). The plates were checked daily using phase contrast microscopy until the colonies (usually consisting of 100 to 200 cells) were large enough to be selected by their morphologic characteristics and uniformity. Glass cloning cylinders (Bellco Glass Inc., Vineland, NJ) were attached to the surface of the dish and around the clone with sterilized high vacuum grease (Dow Coming Corporation, Midland, MI). The surface of the dish isolated by the

INTRODUCTION Endothelial cells (EC) from different organs as well as from large and small vessels of the same organ have specialized properties (3,9,12). There may be organ-specific antigens on the capillary EC membrane (1) as well as unique profiles of cell membrane glycoproteins (2). Regional differences in the negative surface charge and lectin-binding domains are also evident in different microvaseular beds (i 8). The functional significance of the cell surface heterogeneity is suggested by the observation that tumor cells are able to bind preferentially to mierovascular EC (14). There may be other, but poorly understood, functional differences in EC; for example, in the monolayer barrier function (3,9,12). The microvessels are the primary sites of transvascular fluid and solute exchange and of endothelial injury in the lungs (10,15). In the present study, we report on a technique for culturing EC from the microvasculature of bovine lungs. To characterize these cells, we have compared the surface glycoprotein patterns of microvessel EC with the mainstem pulmonary artery- and vein-derived EC and the permeability characteristics of cultured EC monolayers. MATERIALSAND METHODS Isolation of lung microvascular endothelium. This method was modified from a technique described for the isolation and culture of microvessel EC from other organs (8). Bovine (Bos taurus) lungs were obtained from calves and transported to the laboratory on ice. Lungs were rinsed with a 10% povidone-iodine solution (Pharmadine, Sherwood Pharmaceutical

711

712

DEL VECCHIO ET AL.

Fro. 1. a, Phase contrast micrograph of a microcolony of cells obtained after collagenase digestion of bovine peripheral tissue lung. Micrograph was taken 2 days after the initiation of the cultures, b, Phase contrast micrograph of cells derived from a microcolony. Cells are homogeneous in appearance and have a typical "cobblestone" morphology characteristic of endothelial cells, c, Incorporation of DiI-Ac-LDL by pulmonary microvessel endothelial cells. Cells showed a characteristic granular pattern reflecting DIl-Ac-LDL incorporation. d, Immunolocahzation of Factor VIII-related antigen showed a typical granular cytoplasmic pattern.

cylinders was washed 3 times with cmf-PBS and the cylinders were filled with a trypsin solution (PET) [0.05% trypsin, 0.02% [ethylene bis (oxyethylenenitrilo)]-tetraacetic acid, and 1% polyvinyl-pyrrohdone prepared in HEPES buffered in 0.9% NaC1]. The action of the enzyme was stopped by the addition of complete medium, and the individual colonies were seeded onto FN-coated plates (surface area of 2 cm 2) in complete medium (1 ml). The wells were fed every 3 to 4 days, and when the cells were confluent they were passaged to 60-mm dishes or 75-cm 2 flasks previously coated with FN. Subsequent passages were made at a ratio of 1:2 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% FBS. Cells were seeded onto glass cover shps for characterization. Isolation and culture of major vessel endothelial cell lines. EC cells from pulmonary artery, pulmonary vein, and dorsal aorta were isolated by standard techniques (11) and cultured in an identical manner to the lung microvascular (DMEM with 20% FBS). All comparisons between cell lines were made using cells within 5 population doublings (PD) of each other (20 to 25 PD). Cell characterization. The ability of EC ceils to incorporate acetylated low density lipoprotein (Ac-LDL) was determined by incubating Ac-LDL labeled with the fluorescent probe, 1,11-dioctadecyl-l-3,3,31,31-tetramethyl-indocarbocyanine perchlorate (DiI-Ac-LDL) (Biomedical Technologies, Inc., Stoughton, MA) (23). The presence of Factor VIII-related antigen was assessed using an antibody against human Factor VIII-related antigen (Atlantic Antibodies, Scarborough, ME) (11). In vitro angiogenesis in three-dimensional collagen I lattices. Three dimensional collagen lattices were prepared by described methods (13). The microvessel EC were passaged at a 1:3 ratio on the coflagen-coated dishes and allowed to grow on the surface of the gels for 2 days. The culture medium was removed and the monulayer was covered with 1 ml of cold collagen mixture. The dishes were incubated at 37 ° C for 30 min to induce gelation. The cultures were viewed and photographed after 2 days. Microscopy of cells growing on polycarbonate filters. For light micros-

copy, EC monolayers were fixed in 2.5% glutaraldehyde in 0.05 M cacodylate buffer, dehydrated in ethanol, and embedded in glycol methacrylate sections (1 #m) cut in a JB-4 microtome, and stained with methylene blueazure II. For transmission electron microscopy, EC monolayers grown on micropore filters were washed in Hanks' balanced salt solution, (pH 7.4), and fixed with electron microscopy-grade 2.5% glutaraldehyde in 0.05 M Caeodylate buffer. The cells were postfixed with 1% osmium tetroxide, en bloc stained with 0.5% uranyl acetate, and dehydrated in ascending concentrations of ethanol. The polymerized monolayers were cut to 0.5-mm 3 blocks, mounted onto tips of polymerized BEEM capsule blocks, and silverto-fight gold (60 to 70-nm thick) sections were cut with a diamond knife. The sections were double-stained with 4% uranyl acetate and 0.4% lead citrate, and viewed with a JEOL 100 CX. Confluent monolayers of EC were processed for scanning electron microscopy according to procedures of Schroeter et al. (16). Protein labeling. EC proteins were metabolically labeled for 18 to 24 h

TABLE 1 CHARACTERISTICS OF TRACER MOLECULES Molecule Sucrose Inulin Ovalhumin Albumin

MolecularWe/ght 342 5000 43 000 69 000

re

Ds7 × 105

5.2 12.0 27.6 36.1

0.721 0.296 0.110 0.093

Key: MW = molecular weight (Daltons), r~ = Stokes-Einstein radius (nm), and t)37 diffusion coefficient at 37 '~ C in water.

713

PULMONARY MICROVASCULAR ENDOTHELIAL CELLS

Fro. 2. a, Phase contrast micrograph of capillary-like "tube" formation by pulmonary microvascular endothehal cells layered between collagen matrices, b, Scanning electron micrograph of microvessel endothelial cells derived from bovine peripheral lung grown on a microporous polycarbonate filter, c, Light micrograph of a cross-section of microporous polycarbonate filter plated with pulmonary microvessel endothelial cells, d, Transmission electron micrograph of two apposed pulmonary microvessel endothehal cells. Cells showed membrane fusion at points where their plasma membranes meet.

at 5 days postconfluence by addition of fresh media containing 10 gCi/ml [aSS]methionine (specific activity > 800 Ci/mM; ICN, Irvine, CA). EC surface proteins were dual labeled using modified cell surface biotinylation procedures (21). The [3zS]-labeled endothelial monolayers were rinsed 3 times with cmf-PBS and biotinylated by addition of 1 ml/dish of 2.5 mM NSH-LC biotin (Pierce, Rockford, IL). The plates were incubated for 30 min with gentle rocking and rinsed 3 times with DMEM to remove label. Cell proteins were solubilized in 500 ~1 of 0.5% NP-40 solubilization buffer containing a protease inhibitor cocktail (2 mM EDTA, 10 #M PMSF, 0.1 mM TPCK, and 5 mM benzamidine). Cell lysates were microfuged for 5 min to remove nuclei, and the protein content was determined by the Lowa'y method. Glycoprotein separation. EC glycoprotcins were isolated by lectin affinity chromatography on a panel of lectin agarose beads concanavalin A (ConA), wheat germ agglutinin (WGA), Ricinis communis, and peanut agglutinin (PNA) (Sigma). Dual-labeled endothelial cell lysates containing 400 ttg protein were loaded onto an equivalent amount of lectin-agarose in microfuge tubes, incubated 2 h with shaking at 4 ° C, rinsed.5 times with 0.1% NP-40 in cmf-PBS, and the bound glycoproteins were eluted by boiling for 3 min in 3 times sodium dodecyl sulfate (SDS)-sample buffer. Protein analysis. EC proteins and glycoproteins were analyzed by SDSpolyacrylamide gel electrophoresis on 5 to 15% gradient gels followed by Western transfer onto nitrocellulose after lectin affinity chromatography (22). Biotinylated proteins were detected.by staining with streplavidin-alkaline phosphatase (BRL Inc., Gaithersburg, MD), and the radiolabeled proteins were detected by autoradiography using KODAK X-O-Mat AR film (Kodak, Rochester, NY). Permeability measurements. The permeability measurements were made as described previously (5,7,17). The chambers for measuring permeability were prepared by gluing a 13-mm-diameter polycarbonate filter with

a pore size of 0.8 ttm to a polystyrene cyhnder with a 13-ram o.d. and a 9-mm i.d. Filters are gelatinized as described (20). The chambers were sterilized by treatment with UV light overnight. The gelatinized filters on the assembled chambers were treated with a 30 #g/ml solution of ovine FN which was aspirated before seeding of the cells. The microvessel endothelial cells were seeded with 0.5 ml of eel! suspension at a density of 2.0 )< 105 cells/ml in DMEM, 20% FBS, 0.01 mM nonessential amino acids, and gentamicin sulfate (50 #g/ml). The chambers were incubated in a CO2 incubator at 5% CO2:95% humidity before use in permeability studies. Permeability across the endothelial barrier was studied by equipping the polystyrene cylinder with a styrofoam flotation collar and allowing the cylinder to float in a 25-ml volume of fluid. Both upper and lower chambers contained 0.5% bovine albumin solution with small amounts of the test molecules in the upper chamber as a tracer molecule. The transendothelial permeability was measured by the movement of tracer molecules through the endothelial monolayer. The clearance was calculated by the slope of the line produced by plotting microliters of fluid cleared per unit time, and the permeability of albumin was calculated by the ratio of clearance to the surface area of the filter. The permeability of the endothelial monolayer alone (PEc) was calculated from the filter permeability (P~) and the permeability of the monolayer on the filter (Pr+Ec) by: 1/PEc = 1 / P r + E c -

1/PF.

(5)

The characteristics of the molecules used in the permeability study are listed in Table 1. 14CSucrose and [carboxyl 14C]inulin-carboxyl (New England NUclear, Boston, MA) was obtained in crystalline form and diluted in saline. 12Sl-aibumin was prepared using the choloramine-T method (4). 125IOvalbumin was prepared using the Bohon-Hunter reagent (24). Free iodine for the iodinated molecules was removed by extensive dialysis against 0.1 M KI in phosphate buffered saline (PBS) (pH 7.4). The labeled tracers

DEL VECCH10 ET AL.

714 Glyeoprotein

A n a l y s i s of B o v i n e P u l m o n a r y E n d o t h e l i a l Derived from Large and Small Vessels

TOTAL

CONA

Cells

PNA

RCA

WGA

d~

18K---~I 12K---~-I I a

b

c

d

a

b

c

d

a

b

c

d

a

b

c

d

a

b

c

d

FIG. 3. Surface glycoproteinanalysis of several bovine endothelialcell lines: a, mainstem pulmonary artery; b, mainstem pulmonary vein; c, pulmonary microvessel;and d, dorsal aorta. Surface proteins were biotinylatedand were selected by lectin affinity(see Methods for details).

were continuouslydialyzed in PBS (pH 7.4) until use. The ratio of free-tobound 12sI was routinely measured before and after each study using the trichloroaceticprotein precipitationmethod, and correction for free 125Iwas made for all data. RESULTS Small colonies of cells were observed 2 days after the initiation of the pulmonary microvessel EC cultures. These cells were homogenous in appearance (Fig. 1 a) and formed a confluent monolayer with a "cobblestone" morphology (Fig. 1 b). Incorporation of DilAc-LDL and staining for Factor VIII-related antigen showed a characteristic granular fluorescence (Fig. 1 c,d). The monolayers of pulmonary microvessel EC cultured on collagen gels reorganized to form capillary-hke tubules within 2 days after a second collagen layer (Fig. 2 a). Formation of tubules allowed differentiation of endothelial cells from mesothelial cells (6). Scanning electron microscopy of the cells grown on polycarbonate filters showed a confluent monolayer of EC (Fig. 2 b). The cell surfaces were free of mycoplasma contamination. Light microscopy Showed a monolayer of EC growing over the entire filter (Fig. 2 c). The electron micrographs showed areas of membrane fusion involving the plasma membranes of adjacent EC (Fig. 2 d). The results from 3SS-labeled EC showed few differences between the large vessel EC and microvessel EC. Marked differences were noted in surface glycoproteins as assessed by the lectin binding profiles. Galactose-containing glycoproteins [molecular weight (M,) 160 to 220 and 40 kDa] binding to Ricinus communis agglutinin (RCA) and Peanut agglutinin (PNA) were selectively expressed in pulmonary microvessel EC (lane C), in contrast to EC from pulmonary artery, pulmonary vein, and dorsal aorta (Fig. 3). Other less

prominent differences were detected in ConA-binding (M r 60 kDa and WGA-binding (M, 160 and 35 kDa) glyeoproteins (Fig. 3). Permeability of pulmonary microvessel EC monolayers to albumin averaged 1.1 _+ 0.1 X 10 -6 cm/s (mean + SEM) (n = 16). EC from the mainstem pulmonary artery had a significantly higher permeability value of 5.1 + 0.3 × 10 -6 cm/s (P < 0.05) (mean + SEM) (n = 20). A comparison was made of the permeability of microvessel EC monolayers and EC derived from mainstem pulmonary artery to tracer molecules of different sizes. The permeability for each tracer molecule (P~c) was corrected for the diffusion coefficient (Dz7) of the tracer molecules (Table 1). The PEc)/D values are plotted as a function of Stokes-Einstein radii of the molecules (Fig. 4). Although pulmonary microvessel EC monolayers were more restrictive to all tracers when compared with pulmonary artery EC monolayers, the degree of restrictiveness of the smaller tracers (i.e. ]4C-sucrose) was greater (P < 0.05) for sucrose and insulin than for ovalbumin and albumin (Fig. 4). DISCUSSION The EC surface glyeoprotein profile as assessed by biotinylation of surface proteins and binding to lectins showed distinct differences in pulmonary microvessel EC compared to EC from the pulmonary artery, pulmonary vein, and the aorta. The most obvious characteristic of pulmonary microvessel EC was greater binding of RCA and PNA (to proteins with Mr of 40 and 160 to 220 kDa), in contrast to the pulmonary artery, pulmonary vein, and aortic EC. The differential lectin binding may be of value in establishing the microvascular origin of pulmonary vascular EC in culture. A poten-

PULMONARY MICROVASCULAR ENDOTHELIAL CELLS 100 • 14C-Sucrose 114C-Inulin • 1251-Ovalbumin t i251-Albumin

80

PA 01.8

MV 5.9

30.5

6.6

17.7 7.1

3.1 2.6

60, PEC/D

PA

lcm-ll 40=

20-

MV

- ~ . ----

8

O-

J0

i 5

I 10

I 15

| 20

I 25

Stokes-Einstein Radius

4 i 30

I 35

l 40

(A)

FIG. 4. Comparison of the molecular sieving properties of monolayers of endothelial cells derived from pulmonary artery (PA) and endothelial cells derived from pulmonary microvessel (MI/). Permeabihty values were divided by the diffusion coefficient at 37 ° C for each molecule and plotted against the Stokes-Einstein radius. Numbers on the top right indicate the actual permeability values for the four tracers in units of cm-~.

tial concern with this approach is that the observed differences in lectin binding may not reflect the in situ characteristics of these cells. Binding of the surface glycoproteins from microvessel EC to RCA and PNA lectins has also been noted in fresh isolates of microvessels from the mouse lung (2), suggesting that the cell surface glycoprotein profile of pulmonary microvessel EC may not be species-specific. Whether the glycoproteins observed in pulmonary microvessel E C are present in the pulmonary microcirculation will require further studies. We also compared functional differences between pulmonary microvessel and mainstem pulmonary vessel EC in culture by assessing the barrier function of the EC monolayers. EC identified as being from pulmonary microvessels were twofold more restrictive to albumin compared to EC isolated from the pulmonary artery. Interestingly, the pulmonary microvessel EC monolayers were 15-fold more restrictive to sucrose (Mr = 342 Da) as compared to albumin (Mr = 69 kDa) than the pulmonary artery EC monolayer. This finding indicates a marked phenotypic difference in the EC barrier function of cultured monolayers from pulmonary microvessels and pulmonary artery. Inasmuch as sucrose does not enter cells and is transported via paracellular routes, the lower sucrose permeability in microvessel EC suggests that these cells form more restrictive interjunctional complexes and do not have the paracellular pathways or gaps noted in cultured EC monolayers derived from the pulmonary artery (5,19). In summary, we describe a method for culturing and characterizing bovine pulmonary microvessel EC. The pulmonary microvessel EC have a unique profile of cell surface glycoproteins different from the pulmonary artery, pulmonary vein, and aortic EC. Moreover, the cultured pulmonary microvessel EC monolayers were more restrictive than pulmonary artery EC monolayers, particularly for the small molecular weight paracellular tracer, sucrose, suggesting that the cultured pulmonary microvessel EC monolayer forms complex intercellular junction. RrVE~NC~ 1. Auerbaeh, R.; Alby, L.; Morrissey, L. W., et al. Expression of organspecific antigens on capillary endothelial cells. Microvasc. Res. 29:401-411; 1985.

715

2. Belloni, P. N.; Nicolson, G. L. Differential expression of cell surface glycoproteins on various organ-derived microvascular endothelia and endothelial cell cultures. J. Cell. Physiol. 136:398-410; 1988. 3. Belloni, P. N.; Tressler, R. J. Microvascular endothelial cell heterogeneity: interactions with leukocytes and tumor cells. Cancer Metastasis Rev. 8:353-389; 1989. 4. Bocci, V. Efficient labelling of' serum proteins with ]35] using chloramine-T. Int. J. Appl. Radiat. 15:449-456; 1964. 5. Cooper, J. A.; Del Vecchio, P. J.; Minnear, F. L., et aL Measurement of albumin permeability across endothelial monolayers in vitro. J. Appl. Physiol. 62:1076-1083; 1987. 6. Chung-Welch, N.; Patton, W. F.; Ameia Yen-Patton, G. P., et at. Phenotypic comparison between mesothelial and microvascular endothelial cell lineages using conventional endothelial cell markers, cytoskeletal protein markers and in vitro assays of angiogenie potential. Differentiation 42:44-53; 1989. 7, Del Vecchio, P. J.; Siflinger-Birnboim, A.; Shepard, J. M., et al. Endothelial monolayer permeability to macromolecules. Fed. Proc. 46:2511; 1987. 8. Folkman, J.; Handenschild, C. C.; Zetter, B. R. Long-term culture of capillary endothelial cells. Proc. Natl. Acad. Sci. USA 76:52175221; 1979. 9. Gerritsen, M. E. Function heterogeneity of vascular endothelial cells. Biochem. Pharmacol. 36:2701-2711; 1987. 10. Horvath, C. J.; Ferro, T. J.; Jesmok, G., et al. Recombinant tumor necrosis factor increases pulmonary vascular permeability independent of neutrophils. Proc. Natl. Acad. Sci. USA 85:9219-9223; 1988. 11. Jaffe, E. A.; Hoyer, L. W.; Nachman, R. L. Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J. Clin. Invest. 52:2757-2764; 1973. 12. Kumar, S.; West, D. C., Ager, A. Heterogeneity in endothelial cells from large vessels and microvessels. Differentiation 36:57-70; 1987. 13. Montesano, R.; Orci, L.; Vassalli, P. In vitro rapid organization of endothelial cells into capillary-like networks is promoted by collagen matrices. J. Cell. Biol. 97:1648-1652; 1983. 14. Nicolson, G. L.; Winkelhake, J. L. Organ-specificity of bloed-borne tumour metastasis by cell adhesion. Nature 255:230-232; 1975. 15. Ryan, U. S. Metabolic activity of pulmonary endothelium: modulation of structure and function. Ann. Rev. PhysioL 48:263-277; 1986. 16. Schroeter, D.; Spiess, E.; Paweletz, N., et al. A procedure for rupturefree preparation of confluently grown monolayer cells for scanning electron microscopy. J. Electron Microsc. Tech. 1:219-225; 1984. 17. Siflinger-Birnboim, A.; Del Vecchio, P. J.; Cooper, J. A., et al. Molecular sieving characteristics of the cultured endothelial monolayer. J. Cell. Physiol. 132:111-117; 1987. 18. Simionescu, M.; Simionescu, N.; Santoro, F., et al. Differentiated microdomains of the luminal plasmalemma of murine muscle capillaries: segmental variations in young and old animals. J. Cell. Biol. 100:1396-1407; 1985. 19. Schnitzer, J. E.; Siflinger-Birnboim, A.; Del Vecchio, P. J., et al. Segmental differentiation of permeability, protein glyeosylation, and morphology of cultured bovine pulmonary vascular endothelium. J. Physiol. Lond. Submitted; 1992. 20. Taylor, R. F.; Pine, T. H.; Schwartz, S. M., eta!. Neutrophil-endothehal monolayers grown on micropore filters. J. Clin. Invest. 67:584587; 1981. 21. Tomasovic, S. P.; Simonette, R. A.; Wolf, D. A., et al. Co-isolation of heat stress and cytoskeletal proteins with plasma membranes. Int. J, Hyperthermia 5:173-190; 1989. 22. Towbin, H.; Staehelin, T.; Gordon, J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76:4350-4354; 1979. 23. Voyta, J. C.; Via, D. P.; Butterfield, C. E., et al. Identification and isolation of endothelial cells based on their increased uptake of acetylated-low density lipoprotein. J. Cell. Biol. 99:2034-2040; 1984. 24. Williams, M. R.; McBride, D. Problems with radiolabelling egg albumin for the Farr ammonium sulfate technique for antibody estimation. J. Immunol. Methods 15:315-323; 1977.