Establishment of a Pure Vascular Endothelial Cell Line from Human Placenta

Placenta (2000), 21, 325–336 doi:10.1053/plac.1999.0492, available online at http://www.idealibrary.com on

Establishment of a Pure Vascular Endothelial Cell Line from Human Placenta V. V. Jingaa,c, A. Gafencua, F. Antohea, E. Constantinescua, C. Heltianua, M. Raicua, I. Manolescua, W. Hunzikerb and M. Simionescua a

Institute of Cellular Biology and Pathology ‘N. Simionescu’, Bucharest, Romania Institute of Biochemistry, University Lausanne, Switzerland Paper accepted 30 November 1999

b

Endothelial cells (EC) from various sectors of the circulatory system have distinct characteristics, some of which have only been identified in cultures upon their isolation from specific organs or tissues. Cultured vascular EC, derived from the human placenta (HPEC), may be helpful for studying their specific function in the fetoplacental unit, such as in the control of maternofetal traffic. In this paper we report an improved method for isolation, purification and culture of HPEC, that implies an enzymatic perfusion of the term placenta, followed by separation of resulting cells on a Percoll density gradient. The inoculated starting suspension was purified by a two-step selection procedure, based on differential trypsinization, leading to a pure population of about 8107 cells/placenta, with 2.7–3.4 population doublings. The average population doubling time during eight passages was 60–65 h and the life span of HPEC was approximately 45–50 population doublings. The cell morphology at optical and electron microscopical level revealed a good differentiation of HPEC, which were endowed with numerous plasmalemmal vesicles (caveolae) and Weibel–Palade bodies. The transendothelial electrical resistance of the HPEC monolayer varied between 22 and 52 Ohm/cm2. The cultures were mycoplasma free, as revealed by fluorescence microscopy using DNA dyes and the polymerase chain reaction (PCR). The negative immunofluorescent reaction for keratin confirmed that the HPEC were not contaminated with either type of placenta cells, as syncytiotrophoblast. Cultured HPEC demonstrated a strong reaction for von Willebrand factor antigen (by fluorescence microscopy), took up AcLDL-DiI and expressed active angiotensin converting enzyme. These characteristics substantiate the endothelial nature of cultured cells. The interactions with different lectins (BS-I, SBA, RCA, UEA and WGA) assessed by fluorescence microscopy and blotting reveal a strong reaction of HPEC with UEA and a negligible reaction with BS-I lectin. WGA lectin displayed a marked fluorescence staining in subconfluent HPEC, and at the level of intracellular clefts in post-confluent cultures. In conclusion: (i) we have obtained a pure line of cultured EC originating from the human placental venous side of the circulatory tree; (ii) the cells have the general characteristics and markers ascribed to EC; (iii) as opposed to large human placental vessels, HPEC do not react to BS-I lectin and, unlike human umbilical vein EC, have a much higher proliferation rate and a long lifespan; (iv) HPEC expressed a characteristic glycosylated coat particularly rich in
--fucose and -GlcNAc containing glycocompounds.  2000 Harcourt Publishers Ltd Placenta (2000), 21, 325–336

INTRODUCTION The vascular endothelium plays a key role in blood–tissue exchanges, providing a compatible container for blood cells, and monitoring body homeostasis. Conclusive data support an active role for endothelium in pathological conditions such as inflammation, immune reactions, atherosclerosis, thrombosis, diabetes, viral infections, bleeding disorders, neoplasia and metastasis. The ubiquitous endothelial cell (EC) is morphoc To whom correspondence should be addressed at: Institute of Cellular Biology and Pathology ‘N. Simionescu’, 8, B.P. Hasdeu Street, sect. 5, P.O. Box 35-14, Bucharest, Romania. Tel: (401) 411 53 10; Fax: (401) 411 11 43; E-mail: jingav@simionescu. instcellbiopath.ro

0143–4004/00/040325+12 $35.00/0

logically and physiologically highly differentiated, as a result of functional adaptation to local needs (for reviews, see Simionescu and Simionescu, 1988, 1992). Large vessel and microvascular EC have distinct characteristics, some of which were identified upon isolation and culturing of cells from specific organs or tissues. Placental vasculature represents a distinct territory, highly specialized in structure and function. Culturing EC derived from the vasculature of human placenta may provide a system for studying the role of these cells in the complex activities of the placenta, at a cellular and molecular level. Among the five structural components of the placental barrier (i.e. syncytium, trophoblastic basal lamina, connective tissue layer, endothelial basal lamina and endothelium) the regulation of transfer between the two bloodstreams  2000 Harcourt Publishers Ltd

326

was considered to be exerted principally by the syncytial trophoblast (Padykula, 1988; Bloxam, Bax and Bax, 1997). Whilst the architecture of the human placenta may favour an extensive and intimate association between maternal and fetal circulation, it also leads to considerable risks for the vertical transmission of pathogens. Besides the trophoblastic layer (transiently undergoing structural defects), the trophoblastic basement membrane and resident macrophages, the capillary endothelial cells and their junctional complexes may also provide defence mechanisms (Burton and Watson, 1997). Malek et al. (1997) have studied the transport of macromolecules in perfusion experiments of isolated cotyledons from human term placenta. Compared with other proteins, such as BSA and IgA, they identified an increased transfer of IgGs from the maternal to the fetal side, suggesting a specific transport mechanism. However, the experimental model could not discriminate between specific involvements of the various morphological structures of placental barrier. In recent years, some attempts have been made to culture EC from different segments of the placenta vasculature (Lang, Dohr and Desoye, 1993; Leach et al., 1993; Leach et al., 1994; Schu¨ tz and Friedl, 1996; Kacemi et al., 1996; Schu¨ tz, Teifel and Friedl, 1997). In this paper we report an improved method for isolation, purification and culture of vascular EC from human term placenta. The procedure involves enzymatic perfusion of the placenta followed by separation of the resulting cellular components on a Percoll density gradient. A two-step selection method using an inoculated starting material produced an EC line featuring good proliferation and differentiation, and preserving all the in vivo characteristics of these cells. The culture of placental vascular EC may aid understanding of their specific role in the control of maternofetal traffic and in the general functioning of the fetoplacental unit. MATERIALS AND METHODS Isolation and cultivation of endothelial cells from human placenta Throughout these experiments, aseptic techniques, sterilized materials and solutions were used and safety guidelines recommended by the Tissue Culture Association for research laboratories that use human tissue were applied; the reagents were cell culture tested (cct) grade and the water was endotoxin free, having an electrical resistivity of 18.2 mOhm. After informed consent, placentae were obtained by vaginal delivery or caesarean sections from women with uncomplicated full-term pregnancies. Immediately after collection, the loose chorionic and amniotic sacs and decidua were removed and the placenta was extensively washed with Ca + + and Mg + + free cold phosphate buffered saline (136.9 m NaCl, 2.7 m KCl, 8.1 m Na2HPO4, 1.5 m KH2PO4, pH=7.3, all from Sigma Chemical Co., St Louis, MO, USA) supplemented with 15 000 U.I./ l heparin (Biofarm, Bucharest, Romania), 1 ml/l of 4 per cent

Placenta (2000), Vol. 21

papaverin, w/v (Biofarm, Bucharest, Romania), 300 U.I./ml penicillin (Antibiotics, Iasi, Romania), 300 g/ml streptomycin (Antibiotics, Iasi, Romania) and 150 g/ml neomycin (Sigma) (PBS2
). The organ was placed in a glass dish (30 cm diameter) containing approx 300 ml cold PBS2
, the umbilical cord was cut to approx 4 cm and a canula was introduced into the umbilical vein and pushed up to the beginning of branching of the vessel. A ligature was placed at the tip of the canula to prevent umbilical vein cells coming into contact with the perfusate, including the enzyme solution. The canula was connected via a silicon tube to a peristaltic pump (MTA Peripump electronic, type 5186, Hungary). Care had to be taken to avoid an air embolism in the venous circulation of the placenta. While perfusing the organ at a slow rate (5 ml/min) with Dulbecco’s Modified Eagle’s Medium (DMEM, Sigma), supplemented with PBS2
(DMEM), at room temperature, the basal plate of the placenta (side of microvascular bed) was either roughened with a scalpel (Schu¨ tz and Friedl, 1996) or cut at the very surface of the cotyledons with fine scissors. Thus the microvascular bed of the stem villi, intermediate and terminal villi in the close vicinity of the basal plate, mostly small venules and capillaries, as opened. The perfusion was considered successful when the perfusate appeared diffusely distributed on this surface; some unperfused cotyledons were also observed. Finally, the placenta was placed, with the microvascular bed facing down, on nylon gauze mounted on the upper side of a siliconized glass funnel (25 cm diameter). The perfusion continued at a higher rate (40–50 ml/min) to a total of 2 l DMEM. The EC were dislodged by perfusing 600 ml of 0.125 per cent (w/v) trypsin 1 : 250 (from bovine pancreas, Sigma) and 0.25 m EDTA (Sigma, USA) in PBS2
, at a rate of 10 ml/min. The first 50 ml of the perfusate was discarded, and the remainder collected sequentially in 12 chilled centrifuge tubes (50 ml each) containing 5 ml of fetal calf serum (FCS, Gibco). The cells were sedimented by centrifugation at 160 g at 4C for 10 min. The 12 sediments were resuspended in a small quantity of DMEM, amalgamated four by four in three 10 ml tubes, maintaining the sequence of harvest, and recentrifuged in the same conditions. The cellular material obtained in each pellet was layered on a Percoll (Sigma, cct) preformed gradient made up of a 9 : 1 (v/v) mixture of Percoll and DMEM (10) and further diluted with DMEM (1) to 55 per cent Percoll (v/v), having a final osmolality of 300 mOsm. A continuous gradient was obtained by centrifugation for 30 min in 7.2 ml Percoll solution contained in 10 ml centrifuge tubes, at 15 000 g, at 4C. Pelleted cells were placed on top of the Percoll gradient and centrifuged at 650 g for 10 min at 4C. After centrifugation, a white band was found 2–3 mm beneath the surface of the top of the Percoll gradient containing sheets or isolated nucleated cells and small vesicles (probably resulting from damaged cells); at the bottom of the tube, a red pellet of erythrocytes was detected (phase contrast microscopy). The white bands from each tube were collected separately in three centrifuge tubes containing 10 ml MCDB 131 medium (Sigma, USA) supplemented with 0.25 m N-acetyl--cysteine, 12.5 

Jinga et al.: Placental Vascular Endothelial Cell Line

-mercaptoethanol, 1 g/ml hydrocortisone, 1.25 g/ml heparin, 2.8 g/ml insulin and 70 g/ml endothelial cell growth supplement from bovine neural tissue, all from Sigma. This culture medium is similar to that used by Schu¨ tz and Friedl (1996) except that epidermal growth factor, crude fibroblast growth factor and human placental cell-conditioned culture medium were omitted, and endothelial growth supplement was added; this is because our preliminary studies showed a better reproducibility with this nutrient medium. The completed medium (MCDB 131c) was supplemented with 15 per cent FCS (v/v) and penicillin (200 U.I./ml), streptomycin (200 g/ml) and neomycin (300 g/ml). The cells were washed with the medium twice by centrifugation at 160 g, at 4C, for 10 min. Each sediment was then resuspended in 4 ml nutrient medium consisting of MCDB 131c with 15 per cent FCS (v/v), penicillin (100 U.I./ml) and streptomycin (100 g/ml), seeded on three plastic Petri dishes (5 cm diameter) and kept at 37C in 3 per cent CO2 in air (v/v) and relative humidity over 95 per cent. Subcultivation and selection. The subcultures were obtained upon exposure to 0.125 per cent trypsin (w/v) and 0.25 m EDTA in PBS2
and centrifugation at 160 g, for 10 min, at 4C. Each pellet was suspended in nutrient medium, plated on plastic or borosilicate glass (Schott-Duran, Germany) Petri dishes and incubated in the conditions mentioned above. The nutrient medium was changed every 3 days. To obtain pure human placental endothelial cells (HPEC) from primary cultures (occasionally contaminated with fibroblast-like cells, possibly pericytes or/and vascular smooth muscle cells), the selection procedure applied was based on the observation that during trypsinization EC became detached from the substrate before cellular contaminants. A similar observation was reported by Shepro and Morel (1993). Consequently, after the primary cultures reached confluence (8–10 days) each dish was trypsinized sequentially, at 1 min intervals for 10 min, and the 10 harvested aliquots per dish were processed separately and plated on 10 dishes (5 cm in diameter). After 7–8 days, the subcultures were trypsinized under microscopic control; the fibroblast-like cells were not detached from the substrate whereas the EC were harvested and inoculated on Petri dishes at a density of 1.2105 cells/cm2. During and after the selection procedure, the cultures were examined daily using an inverted microscope. When possible, the well delimited islands of contaminant cells were marked and removed, using the tip of a Pasteur pipette. The culture dishes that exhibited a heavy contamination were discarded. Freezing procedure. Stock cultures were prepared by freezing the cells in MCDB 131c medium with 15 per cent FCS and 10 per cent dimethyl sulphoxide (v/v) (Sigma, cct) at a freezing rate of 1C/min. in aliquots of 106 cells/ml/cryogenic tube. Two-chamber system. HPEC were grown on transparent filter inserts (0.4 m pore size) in a Transwell-Col culture chamber from Costar (Van Nuys, CA, USA). The tightness of cell

327

monolayers was assessed by measuring the transendothelial electrical resistance with a Millipore Millicell Electrical Resistance System. Mycoplasma screening. HPEC culture were screened periodically for mycoplasma contamination by fluorescence microscopy, using the selective DNA dye Hoechst 33258 (Sigma) (filter cube: UV-1A) (Chen, 1977; Freiberg and Masover, 1990), and by PCR with the Mycoplasma PCR primer set (Stratagene Cloning Systems, La Jolla, CA, USA). The contaminated material was discarded. In some cases bovine aortic endothelial cells (BAEC) were cultured, as in Jinga, Bogdan and Fruchter (1986) as external controls; in addition human aortic smooth muscle cells (HASMC) obtained from explants were cultured in the same conditions as BAEC. Characterization of HPEC Morphological studies. For routine and phase contrast observations an inverted microscope Telaval-Zeiss (Jena, Germany) and a photomicroscope Docuval-Zeiss (Jena, Germany) were used, respectively. For transmission electron microscopy (TEM), HPEC were grown on 3.5 cm plastic dishes and processed as described in Jinga, Bogdan and Fruchter (1986). All buffer solutions had a pH 7.4 and an osmolality of 300 mOsm. Briefly, the cultures were washed twice (30 sec) with 75 m sodium cacodylate (Serva)–HCl (Sigma) buffer, supplemented with 3 per cent (w/v) sucrose (Merck) (SCB) and fixed for 5 min in 2.5 per cent glutaraldehyde (Merck) in SCB. After two brief washings with SCB, the cultures were post-fixed in 2 per cent (w/v) osmium tetraoxide (Sigma) in 150 m sodium cacodylate–HCl buffer (CB) for 10 min, at 4C, then washed twice for 2 min in CB followed by 1 per cent (w/v) tannic acid (Mallinkrodt) for 7 min (Simionescu and Simionescu, 1976) and two rinses (2 min) in 1 per cent (w/v) sodium sulphate (Sigma) in 100 m cacodylate–HCl buffer. The cultures were then rapidly dehydrated in increasing concentrations of ethanol (Merck): 70 per cent (v/v) (215 sec), 95 per cent (15 sec) and 100 per cent (35 sec) and infiltrated with 1 : 1 (v/v) Epon (Polysciences, USA) ethanol (100 per cent) for 30 min. After the resin mixture was discarded, the cells were kept for 10 min, at 37C (for ethanol evaporation), followed by embedding in Epon 812, and polymerization at room temperature for 1 h, at 37C for the next 4 h and at 55C for 2 days. The embedded cultured cells were separated from the plastic surface by rapid immersion in liquid nitrogen. Thin sections obtained on a Reichert OmU3 ultramicrotome were stained with 7.5 per cent (w/v) uranyl acetate (Polysciences, USA) for 10 min with 0.4 per cent (w/v) lead citrate (Merck, Germany) for 1.5 min and examined with a Philips 201C and a Philips 400 (Holland) transmission electron microscope. Expression of von Willebrand factor antigen. HPEC cultured on glass coverslips were rinsed three times with DMEM, fixed in

328

Placenta (2000), Vol. 21

Figure 1. Human placental endothelial cells in culture, 3rd passage. Phase contrast microscopy. (A) Proliferative culture, 36 h after seeding. Note the elongated shape of the cells. Bar: 77 m. (B) Preconfluent stage of the culture. Bar: 100 m. (C) Post-confluent monolayer exhibiting convex polygonal cells. Bar: 111 m. (D) Higher magnification of polygonal cells constituting the monolayer, with prominent nuclei and nucleoli. Bar: 22 m.

cold acetone (
20C) for 5 min and air-dried at room temperature (RT). Since the secondary antibody was goat IgG, normal goat serum was used for all the dilutions to prevent non-specific reactions. The cells were incubated in normal goat serum diluted 1 : 20 in 0.01  PBS, pH 7.4 (buffer T), for 15 min. Rabbit IgG anti-human von Willebrand factor antigen (Sigma) was diluted 1 : 1000 in the buffer T; normal rabbit IgG was used as a control. The glass coverslips with cultured cells (upside down) were exposed to antibody in a moisture chamber, at 37C, for 1 h. After extensive washing with buffer T diluted 1 : 10 (315 min, with gentle shaking) to remove the unspecifically bound antibodies, the cells were incubated in the same conditions, with the goat anti-rabbit affinity purified IgG labelled with TRITC, 1 : 30 dilution (Sigma). After 1 h, at 37C, the coverslips were washed thoroughly with PBS (315 min), fixed in 2 per cent paraformaldehyde (J.B. EM Services, Inc., Quebec, Canada) in PBS for 10 min, rinsed in double distilled water (DDW) and mounted in a drop of SlowFade (Molecular Probes, Eugene, OR, USA). The cells were examined with a fluorescence microscope (Microphot SA Nikon, Tokyo, Japan) equipped with a filter combination G-1B for rhodamine and photographed on a T-MAX 400ASA Kodak film. Detection of keratin filaments. The same method as for von Willebrand factor antigen was used, except that the primary

antibody was monoclonal ascites anti-human pan-cytokeratin (Sigma) and the secondary antibody was goat affinity purified IgG anti-mouse (Fab specific) coupled with TRITC (Sigma). Both antibodies were diluted 1 : 100 in buffer T. Uptake of acetylated low density lipoproteins (AcLDL). AcLDL coupled with fluorescent 1,1 -dioctadecyl-3,3,3 ,3 -tetramethyl-indocarbocyanide perchlorate (Molecular Probes, USA) (AcLDL-DiI) was prepared as described in Voyta et al. (1984). Confluent HPEC on glass coverslips were washed with PBS containing 1.2 m CaCl2 and 0.5 m MgCl2 · H2O (Sigma, cct), incubated with AcLDL-DiI (10 mg/ml) for 1 h and examined with the fluorescence microscope (filtercombination G-1B). Controls consisted of similar processed cultures, except that AcLDL-DiI was omitted from the incubation medium. Lectin studies. The lectins used (all from Sigma) were: Bandeiraea simplicifolia (BS-I), Glycine max (SBA), Ricinus communis (RCA), Triticum vulgare (WGA) and Ulex europaeus (UEA). Lectin fluorescence microscopy. For this approach HPEC or BAEC (the latter as an external control) grown on coverslips were sequentially: (a) washed with DMEM (21 min) at 37C; (b) fixed for 30 min at room temperature (RT) with 3.5 per cent formaldehyde (Sigma) in PBS; (c) washed with

Jinga et al.: Placental Vascular Endothelial Cell Line

329

DMEM at RT (31 min); (d) quenched with 1.5 per cent (w/v) fish skin gelatin (Sigma) in PBS (PBSG); (e) incubated for 1 h at RT with a given biotinylated lectin (50 g/ml) in PBSG; (f) washed in PBSG (21 min); (g) incubated for 1 h, at RT, with 1 g/ml Streptavidin–Texas Red (Amersham, England) in PBSG; (h) washed in PBS (31 min); and finally (i) mounted and examined with the fluorescence microscope equipped with the filter combination G-1B and photographed on T-MAX 400 ASA Kodak film. Lectin blotting. Confluent monolayers of HPEC were washed with DMEM at 37C (31 min) and then at 4C (21 min). The cells were lysed by exposure to the cold solubilization buffer containing 1.17  Tris-HCl (Sigma), pH 6.8, 3 per cent (w/v) sodium dodecyl sulphate (SDS) (Sigma), 1.2 per cent (v/v) -mercaptoethanol (Sigma), 2  urea (Sigma) and 3 m EDTA (Sigma) in DDW. Cell nuclei and debris were pelleted by centrifugation at 1000 g, for 10 min, and the supernatant (the cell lysate) was subjected to preparative SDS-PAGE (acrylamide, bisacrylamide from Sigma) (200 g protein/lane) followed by electrotransfer onto nitrocellulose filters (BA-85, 0.45 m from Schleicher and Schuell Inc., England) using the following transfer buffer: 25 m Tris, 190 m glycine and 20 per cent methanol (v/v) in DDW, with the pH adjusted to 7.8–8.3 with glacial acetic acid (all from Sigma). The filters were stained briefly with Ponceau S (Sigma), washed in DDW and dried. Strips cut from the filter were: (a) quenched for 1h with PBS + (containing 1 m CaCl2, 0.05 per cent Nonidet P-40 and 1 per cent (w/v) fish skin gelatin—all from Sigma); (b) incubated for 1 h with 0.2 g/ml biotinylated lectin in PBS + ; (c) washed with PBS + (310 min); (d) incubated for 1 h with 1 g/ml of streptavidin conjugated to alkaline phosphatase (Amersham, England) in PBS + and then (e) washed with PBS-Ca (PBS containing only 1 m CaCl2) (310 min). The strips were incubated with the BCIP/NBT phosphatase substrate system kit (5-bromo 4-chloro-3-indolyl-phosphate and Nitroblue tetrazolium in 0.1  Tris buffer solution) as per manufacturer’s instructions (Amersham, England). The reaction was stopped by placing the strips back into PBS for 1 min, followed by air drying. As controls strips were also processed as above in the absence of the lectins. Demonstration of angiotensin converting enzyme Angiotensin converting enzyme (ACE) activity. The activity of angiotensin converting enzyme (ACE) was assessed in the cellular pellet and the culture medium by using hippuryl-histidyl--leucine (Sigma) as a substrate and cyanuric chloride as a colour reactive (Sigma) according to Schnaith et al. (1994). Expression of angiotensin converting enzyme. ACE expression was determined by immunoprecipitation of metabolically labelled cells. Briefly, confluent HPEC and HASMC (as a negative external control) were washed three times with PBS and starved in methionine free DMEM, containing 15 per cent dialysed FCS (labelling medium) for 30 min, at 37C. The

Figure 2. Electron microscopic aspect of confluent placental endothelial cells (EC). (A) Two EC with long overlaps and intercellular junctions. ij, intercellular junction. Bar: 0.59 m. (B) Basal region of an EC with many vesicles (v) opened to the extracellular space. A coated vesicle (cv) could also be noted. The extracellular space is characterized by filamentous matrix components. pm: plasmalemma; ecm: extracellular matrix. Bar: 0.24 m. (C) The juxtanuclear area of an endothelial cytoplasm rich in cellular organelles. The Golgi complex composed of groups of cisternae appears in numerous copies and some Golgi vesicles are budding from the stacks. A large number of mitochondria and rough endoplasmic reticulum is in close proximity to the Golgi complex. At the basal site, many plasmalemmal vesicles could be seen. Go: Golgi complex; c: centriole; GoV: Golgi vesicle; m: mitochondria; rer: rough endoplasmic reticulum; v: plasmalemmal vesicle; N: nucleus. Bar: 0.88 m.

cells were exposed for 16 h, at 37C, to fresh labelling medium containing 90 Ci/ml [35S]--methionine (TRAN 35S Label) (ICN Pharmaceutical Inc., CA, USA). After the labelling medium was removed, the monolayers cooled on ice were washed three times with PBS, separated from plates with a rubber scraper and further solubilized in lysis buffer containing 50 m Tris HCl buffer, pH 7.5, supplemented with

330

Figure 3. (A) Detail from the Golgi complex region, showing the well organized Golgi stacks, running parallel to each other surrounded by numerous Golgi vesicles and mitochondria. GoS: Golgi stacks; GoV: Golgi vesicles; m: mitochondria; N: nucleus. Bar: 0.37 m. (B) Cytoplasmic area containing high-developed rough endoplasmic reticulum. The tubular cisternae with attached ribosomes have a relatively dense content and are positioned next to the Golgi complex. rer: rough endoplasmic reticulum; Go: Golgi complex; m: mitochondria. Bar: 0.78 m.

150 m NaCl, 2 m EDTA · 2Na, 1 per cent Triton X-100 and 35 g/ml PMSF for 30 min, at 4C. After 3 min centrifugation, at 13 000 g, aliquots of post-nuclear supernatant were immunoprecipitated by sequential incubation with goat antibody against rabbit pulmonary ACE for 18 h, at 4C, then with anti-rabbit IgG biotinylated species specific F(ab )2 fragment as a secondary antibody for 5 h, followed by a suspension of streptavidin–agarose beads for 18 h. Extensively washed agarose bead immunoprecipitates were boiled for 5 min in Laemmli solubilizing buffer (Laemmli, 1970) supplemented with -mercaptoethanol, and then analysed by SDS-PAGE (5–15 per cent polyacrylamide gradient). Coomassie stained gels were dried and exposed on Kodak X-Omat film for 3–10 days at
70C. RESULTS Phase contrast microscopy Immediately after the beginning of cellular growth in culture and during the period of logarithmic proliferation, HPEC

Placenta (2000), Vol. 21

Figure 4. Distal region of the cytoplasm containing a large number of Weibel–Palade bodies. (A) In longitudinal section they have a long, rod-like shape, with tubular content. Bar: 0.36 m. (B) In cross-section, their microtubular internal substructure is better identified. W-P: Weibel–Palade body; v: vesicle; ecm: extracellular matrix. Bar: 0.21 m.

exhibited an elongated morphology; 10–12 days after the inoculation, the primary cultures reached confluence, and consisted mainly of epithelial-like cells. With increased cellular density, in post-confluent cultures, the cells acquired a convexpolygonal shape, 20/40 m in size, with prominent nuclei containing one or more nucleoli (Figure 1). This phenomenon has previously been reported in other human vascular EC cultures (Gallery et al., 1991; Kraling and Bischoff, 1998). At this stage, occasionally sparse or associated fibroblast-like cells were identified. After the selective two-step trypsinization and mechanical removal of contaminant cells, a pure population of epithelial-like cells was obtained; further characterization confirmed the endothelial nature of these cells. The selected population produced at confluence 8–12104 cells/cm2. From a ‘productive’ placenta it was possible to obtain a cryoconserved stock of about 8107 cells, with 2.7–3.4 population doublings. The rate of recovery after freezing was in the range of 85–90 per cent. The average population doubling time during eight passages was 60–65 h. After the 12th passage progressive

Jinga et al.: Placental Vascular Endothelial Cell Line

331

morphological alterations, such as increased cell size, overlapping of cells and cytoplasmic vacuolization, were noted; in addition there was an extended average population doubling time and a decrease in saturation densities, leading to the end of culture after approximately 45–50 population doublings. The transendothelial electrical resistance of HPEC cultures (grown on filter inserts in Transwell-Col culture chamber) varied between 22 and 52 Ohm/cm2; these figures were in the range reported for cultured bovine aortic endothelial cells (Rutten, Hoover and Karnovsky, 1987). Electron microscopy The ultrastructural study of confluent cultured HPEC revealed the monolayer arrangement and the specific characteristics of a continuous vascular endothelium. Adjacent cells were connected to each other by long extensions, bridged by intercellular junctions [Figure 2(A)], indicating the existence of mechanical links between apposed cells and their function in the control of paracellular permeability. HPEC were endowed with numerous plasmalemmal vesicles (caveolae) characteristic of vascular EC, which appeared at both apical and basal surfaces. Figure 2(B) shows an area of the basal front of a cell with a coated vesicle situated in close proximity to the basal plasmalemma and some caveolae opened to the extracellular space rich in fibrillar matrix components. Within the juxtanuclear area, the cell cytoplasm contained a large number of organelles [Figure 2(C)]: multiple copies of Golgi complexes, numerous mitochondria and rich rough endoplasmic reticulum, indicating an intense secretory activity. At higher magnification, the Golgi complex presented characteristic, well organized stacks, running in parallel and surrounded by numerous vesicles and mitochondria [Figure 3(A)]. In close proximity to the Golgi complex the endoplasmic reticulum cisternae had a relatively dense content and were lined by attached ribosomes [Figure 3(B)]. Distal to the nucleus, the cell cytoplasm contained Weibel– Palade bodies, organelles specific to the EC (Weibel and Palade, 1964). As shown in Figure 4(A), they appeared as long, rod-like shape, having an internal tubular content that can be easily identified in cross-section [Figure 4(B)]. In the same cell monolayer we found two different cell types, distinguished by ultrastructural morphology: one type had the general pattern already described [Figure 3(B)], with well organized rough endoplasmic reticulum and Golgi complexes, and the other type was characterized by the presence of numerous cytoplasmic free ribosomes (Figure 5). We do not know whether these ultrastructural aspects represent different metabolic stages of the same type of cells, or the culture contains indeed two different cell types.

Figure 5. The second endothelial cell type found in the same monolayer. The cell cytoplasm contains only free ribosomes and no rough endoplasmic reticulum as shown in the first type of cell. Bar=0.2 m.

extensive and uniform reaction demonstrated the endothelial nature of the cells (Hoyer, de los Santos and Hoyer, 1973) and the purity of the culture. At a higher magnification, it appeared evident that the fluorescent antibody labelled rod-like structures spread throughout the cytoplasm [Figure 6(B)]. These results corroborate well with the reported strong fluorescence labelling for vWf antigen found in microvascular human placental EC obtained from isolated villous microvessels (Kacemi et al., 1996). Immunolabelling for keratin The immunofluorescent reaction of anti-keratin monoclonal antibody was completely negative, indicating that the HPEC culture was not contaminated by epithelial cells. Uptake of acetylated LDL was demonstrated after the incubation of live HPEC in culture with the fluorescent lipoprotein. All the endothelial cells were brightly stained. The dot-like fluorescence appeared throughout the cell cytoplasm but was especially concentrated in the perinuclear region of the cell (Figure 7). Angiotensin converting enzyme activity was present both in cultured HPEC homogenates and in nutrient medium collected from the culture dish after 48 h. As expected, HASMC processed in parallel failed to exhibit any ACE activity. To prove the expression of ACE, homogenates of metabolically labelled HPEC and HASMC with 35S-methionine were interacted with ACE antibody. The resulting immunoprecipitates analysed by SDS-PAGE autoradiography indicated the presence of a polypeptide of approximately 220 kDa in HPEC homogenates, similar in size to ACE, whereas HASMC did not express ACE (Figure 8). Interaction with lectins

Von Willebrand factor antigen All cultured placental endothelial cells were strongly labelled by the von Willebrand factor antibody [Figure 6(A)]. This

The interaction of different lectins (BS-I, SBA, RCA, WGA, UEA) was investigated by fluorescence microscopy (HPEC and BAEC) and lectin blotting (HPEC).

332

Placenta (2000), Vol. 21

Figure 6. Localization of von Willebrand factor antigen in preconfluent cultures of human placental endothelial cells reacted with von Willebrand factor antibody. The fluorescence reaction appears as punctate or rod like structures distributed throughout the cytoplasm of cultured cells (A) that can be better identified at a higher magnification (B). Bar A: 40 m, B: 10 m.

Figure 7. The ability of HPEC to internalize acLDL-Dil. After incubation with acLDL-Dil, HPECs were brightly stained; the fluorescence was distributed as dots located throughout the cells. Bar: 20 m.

The histochemical studies revealed that BS-I lectin reacted weakly, or did not react at all, with HPEC [Figure 9(A)], whereas UEA lectin gave a significantly positive reaction [Figure 9(B)]. Using the same methods, BS-I gave a bright

fluorescence on BAEC [Figure 9(C)] whereas UEA lectin fluorescence was absent [Figure 9(D)]. As stated before, BAECs were used only as external controls, positive for BS-I and negative for UEA lectins (Lang et al., 1994).

Jinga et al.: Placental Vascular Endothelial Cell Line

kDa 205

116 94

66

a

b

Figure 8. Expression of ACE in human placental endothelial cells (HPEC). Metabolically labelled HPEC (lane a) and human aortic smooth muscle cells (lane b) with [35S]--methionine were lysed and treated with goat antibody against rabbit pulmonary ACE antibody. Immunoprecipitates were analysed by SDS-PAGE on 5–15 per cent polyacrylamide gels and fluorography. Positions of molecular weight markers are indicated on the right. Note the presence of ACE in HPEC only (arrowhead). Few polypeptides (in lane a) gave mild cross-reactions with the antiserum.

Sub-confluent HPEC exposed to WGA lectin displayed a marked fluorescence staining, while in post-confluent cultures the lectin stained mainly the intercellular clefts [Figure 10(A, B)]. The controls performed in the absence of biotinylated lectin were negative for all the lectins tested. The lectin blotting results are presented in Table 1. HPEC nitrocellulose strips exposed to BS-I (specific for
-Gal and
-GalNAc), SBA (for
and -GalNAc and
and -Gal) and RCA lectin (specific for
-Gal, -Gal and GalNAc) revealed a similar lectin binding profile; a stronger signal was detected for UEA (specific for
--fucose) and WGA (specific for
-GlcNAc and NeuAc). The results demonstrated that the surface of cultured placental EC preserve an elaborate carbohydrate coat, particularly rich in
--fucose, and
-GlcNAc containing glycocompounds. DISCUSSION Unlike the favourable circumstances provided by large vessels, the isolation of vascular endothelial cells is tedious (Olsen, 1994). The need to investigate the peculiarities of different vascular segments has led to numerous different protocols for isolation and propagation of various human vascular EC (Folkman, Haudenschild and Zetter, 1979; Davison, Bensch and Karasek 1980; Voyta et al., 1984; Jarrell et al., 1986; Jackson et al., 1990; Dorovini-Zis, Pramcya and Bowman, 1991; Fawcett, Harris and Bicknell, 1991; Stansby et al., 1991; Carley, Niedbala and Gerritsen, 1992; Hewett and Murray, 1993; Nishida et al., 1993). Although the perfusion of an organ or tissue vascular bed is generally difficult, in the particular case of the placenta this was feasible because the microvasculature of the stem villi and the anastomosed network of

333

sinusoidal capillaries (that may exceed 50 m in diameter) (Padykula, 1988) represents a rather low resistance pathway. Thus, two basic approaches were previously used for isolating placental vascular endothelium: (1) separation of placental microvessels giving a mixture of pericytes and microvascular EC (Chalier, Kacemi and Olive, 1995). The latter were further separated by magnetic microbeads coated with specific markers for EC (Jackson et al., 1990; Leach et al., 1994; Kacemi at al., 1996); (2) Perfusion of the placenta with enzymatic solutions (Schutz and Friedl, 1996 and this paper). The method presented in this paper has the advantage that fractional harvesting of the dislodged cells during enzymatic perfusion insured a good viability of the cells, and the Percoll gradient and differential trypsinization of the primary and secondary cultures led to a pure population of cultured EC. Some trivial observations about isolation and selection steps of EC from placental are worth mentioning to spare time and effort. The perfusion of the placenta has to be set up immediately after expulsion, before intravascular clotting occurs. The presence of an anticoagulant in the perfusion fluid, the cannulation of the umbilical vein (so as to avoid dislocation of EC from this location) and the thermic balance between the perfusate and the placental tissue (to avoid air embolization) led to successful washing and enzymatic treatment. The quality of the placenta in providing a good starting material, able to propagate in vitro, was of paramount importance. The search for a proper donor (apparently time consuming) was eventually rewarding by saving subsequent use of growth factors, conditioned media, substrate coating materials and by providing prolific and long life cell populations, capable of good differentiation. The primary suspension plausibly contained cells from different segments of the venous front of placental circulation. In our perfusion system the small vessels displayed an important aria, probably supplying the main number of cells. During the establishment period of the culture some subpopulations, endowed with selective advantages, may propagate preferentially. It was rather difficult to precisely locate the origin of our ECs, although subsequent efforts, made to define the purity and homogeneity of HPEC and the structural and biochemical differentiation, suggested small vessel endothelium features of our culture. Von Willebrand factor (vWf) antigen was detected in EC of human placenta by Labarrare et al., 1990. With our method of separation, all HPECs presented a strong labelling for vWf antigen. The specific immunofluorescence was associated with numerous cytoplasmic particles, that had a specific rod-like pattern, different from the usual granular aspect seen in other types of endothelium; they may represent both Weibel–Palade bodies and/or branched endoplasmic reticulum in which vWf was secreted. Electron microscopy confirmed the presence of Weibel–Palade bodies and a well-developed rough endoplasmic reticulum within the cytoplasm. These data corroborate well with the reported strong expression of vWf antigen in cultured EC obtained from isolated human villous microvessels (Kacemi et al., 1996).

334

Placenta (2000), Vol. 21

Figure 9. Histochemical detection of lectin (BS-I, A, C; UEA, B, D) binding glycoproteins expressed on the surface of HPEC (A, B) and BAEC (C, D) in culture. The cells were fixed with 3.5 per cent formaldehyde and incubated with the corresponding biotinylated lectin followed by Streptavidin-Texas Red. Note the positive reaction of UEA on HPEC (B) and of BS-I on BAEC (C). Bar: 20 m.

Among other EC markers, the uptake of AcLDL-Dil and the demonstration of ACE activity and ACE expression in cultured HPEC substantiated their endothelial nature. HPEC did not present any fluorescent reaction for keratin providing confirmation that the cultured cells were not contaminated with other type of placental cells, such as syncytiotrophoblasts, known to contain keratin filaments. The EC lack keratin, but contain vimentin, an intermediate type of cytoskeletal filament. Ultrastructural study revealed a reasonably good differentiation of HPEC. The specific characteristics of a continuous epithelium correlated well with the electrical resistance of the monolayer cultured on transparent filter inserts in twocompartment Costar culture chambers. Apparently, two varieties of EC were present in the same monolayer: a secretory type, featuring well-developed Golgi complex and rough endoplasmic reticulum, and another type which contained

numerous free ribosomes in the cytoplasm. These data are in good agreement with in situ observations. In a morphometric study on human term placental capillary EC, Karimu and Burton (1995) found, in the same capillary cross-sectional profile, two cell types with different organelle distribution patterns failing to detect a specific localization or specialization of these types of endothelium along the length of the capillary wall. An individualization of these two types of cells in culture could help in obtaining information about their significance and function in vivo. The expression of some glycoproteins in cultured HPEC was demonstrated by the results obtained from lectin blotting and lectin fluorescence microscopy. As shown, BS-I lectin staining was weak or absent on HPEC. This corroborates well with the data obtained in situ by Lang et al. (1994) that report the absence of BS-I reaction on the microvasculature of the stem villi of human placenta, showing an increased staining of

Jinga et al.: Placental Vascular Endothelial Cell Line

335

Table 1. Lectin-blotting and fluorescence

Lectin

Lectin Lectin fluorescence blotting Abbreviation HPEC HPEC BAEC

Bandeiraea simplicifolia Glycine max Ricinus communis Triticum vulgare Ulex europaeus

BS-I SBA RCA WGA UEA

 + + ++ ++

 N N ++ ++

++ N N + 

Representation of the interaction of HPEC and BAEC with different lectins; + + to +: a strong to mild signal; : a weak signal when compared to controls; N: experiment not done.

Figure 10. Fluorescence staining of subconfluent (A) and confluent (B) HPEC with biotinylated WGA and Streptavidin-Texas Red. The lectin binding glycoproteins have a different distribution as a function of growing stage of the cells. Bar: A: 40 m; B: 20 m.

the endothelium deeper in the chorionic plate and a decreased reactivity in the stem villi with decreasing vessel size. As expected, a strong reaction on HPEC was obtained with UEA,

which is reported to be a marker for vascular endothelium of human origin (Holtho¨ fer et al., 1982), including ECs from human term placental villi (Leach et al., 1994; Lang et al., 1994). In addition, WGA lectin, known to react strongly with continuous vascular endothelium, both in situ and in culture (Schnitzer, Shen and Palade, 1990), displayed a sharp fluorescence staining on cultured HPEC. The signals obtained using lectin blotting were convergent with those obtained by fluorescence microscopy. Together, these data indicate that by the procedure used (i) we obtained a pure culture of HPEC; (ii) as opposed to large human placental vessels, HPEC do not react with BS-I lectin and unlike cultured large vessel EC, but like HUVEC, they have a high proliferation rate and a significantly longer life span (40–50 population doublings); (iii) the cells preserve in culture the general characteristics and markers ascribed to human EC.

ACKNOWLEDGEMENTS Goat antibody against rabbit pulmonary ACE was a generous gift from Dr R. L. Soffer, Dept. Molec. Biology, Albert Einstein Coll. Med., New York, USA). We are grateful for the excellent assistance of Mrs Marilena Daju and Mrs Mihaela Schean in preparation of the manuscript and figures. This work was supported by Swiss National Science Foundation grant #7RUPJO48567 and Romanian Research and Technology Ministry grant #4103/98.

REFERENCES Bloxam DL, Box BE & Box CMR (1997) Culture of syncytiotrophoblast for the study of human placental transfer. Part II: production, culture and use of syncytiotrophoblast. Placenta, 18, 99–108. Burton GJ & Watson AL (1997) The structure of the human placenta: implications for initiating and defending against virus infections. Rev Med Virol, 7, 219–228. Challier JC, Kacemi A & Olive G (1995) Mixed culture of pericytes and endothelial cells from fetal microvessels of the human placenta. Cell Mol Biol, 41, 233–241. Carley WW, Niedbala MJ & Gerritsen ME (1992) Isolation, cultivation, and partial characterization of microvascular endothelium derived from human lung. Am J Respir Cell Mol Biol, 7, 620–630. Chen TR (1977) In situ detection of mycoplasma contamination in cell cultures by fluorescent Hoechst 33 258 Stain. Exp Cell Res, 104, 255. Davison PM, Bensch K & Karasek MA (1980) Isolation and growth of endothelial cells from the microvessels of the newborn human foreskin in cell culture. J Invest Dermatol, 75, 316–321.

Dorovini-Zis K, Prameya R & Bowman PD (1991) Culture and characterisation of microvascular endothelial cells derived from human brain. Lab Invest, 64, 425–436. Fawcett J, Harris AL & Bicknell R (1991) Isolation and properties in culture of human adrenal capillary endothelial cells. Biochem Biophys Res Commun, 174, 903–908. Folkman J, Haudenschild CC & Zetter BR (1979) Long-term culture of capillary endothelial cells. Proc Natl Acad Sci USA, 76, 5217–5221. Freiberg EF & Masover KG (1990) Mycoplasma detection in cell culture by concomitent use of Bis-benzamide and fluoresceinated antibody. In Vitro Cell and Devel Biol, 26, 585. Gallery ED, Rowe J, Schrieber L & Jackson CJ (1991) Isolation and purification of microvascular endothelium from human decidual tissue in the late phase of pregnancy. Am J Obstet Gynecol, 165, 191–196. Hewett PW & Murray JC (1993) Human lung microvessel endothelial cells: Isolation, culture and characterization. Microvasc Res, 46, 89–102. Holtho¨fer H, Virtanen I, Kariniemi A-L, Hormia M, Linder E & Miettinen A (1982) Ulex europeaus I lectin as a marker for vascular endothelium in human tissues. Lab Invest, 47, 60–66.

336 Hoyer LW, de los Santos RP & Hoyer JR (1973) Antihemophilic factor antigen localization in endothelial cells by immunofluorescent microscopy. J Clin Invest, 52, 2737–2744. Jackson GJ, Garbett PJ, Nissen B & Schrieber L (1990) Binding of human endothelium to Ulex europaeus I-coated dynabeads: application to the isolation of microvascular endothelium. J Cell Sci, 96, 257–262. Jarrell BE, Williams SK, Stokes G, Hubbard FA, Carabasi RA, Koolpe E, Greener D, Pratt K, Moritz MJ, Radomski J & Speicher L (1986) Use of freshly isolated capillary endothelial cells for the immediate establishment of a monolayer on a vascular graft at surgery. Surgery, 100, 392–399. Jinga VV, Bogdan I & Fruchter J (1986) Experimental model for the quantitative estimation of transendothelial transport in vitro: a twocompartment system. Rev Roum Morphol Embryol Physiol, 23, 227–236. Kacemi A, Challier JC, Galtier M & Olive G (1996) Culture of endothelial cells from human placental microvessels. Cell Tissue Res, 283, 183–190. Karimu AL & Burton GJ (1995) Human term placental capillary endothelial specialization. A morphometric study. Placenta, 16, 93–99. Kra¨ ling BM & Bischoff J (1998) A simplified method for growth of human microvascular endothelial cells results in decreased senescence and continued responsiveness to cytokines and growth factors. In Vitro Cell Dev Biol Anim, 34, 308–315. Labarrere CA, Esmon CT, Carson SD & Faulk WP (1990) Concordant expression of tissue factor and class II MHC antigens in human placental endothelium. Placenta, 11, 309–318. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680–691. Lang I, Dohr G & Desoye G (1993) Isolation and culture of fetal vascular endothelial cells derived from human full term placenta. Placenta, 14, A40. Lang I, Hahn T, Dohr G, Skofitsch G & Desoye G (1994) Heterogeneous histochemical reaction pattern of the lectin Bandeiraea Griffonia simplicifolia with blood vessels of human full-term placenta. Cell Tissue Res, 278, 433–438. Leach L, Clark P, Lampugnani MG, Arroyo AG, Dejana E & Firth JA (1993) Immunoelectron characterization of the inter-endothelial junctions of human term placenta. J Cell Sci, 104, 1073–1081. Leach I, Bhasin Y, Clark P & Firth JA (1994) Isolation of endothelial cells from human term placental villi using immunomagnetic beads. Placenta, 15, 335–364. Malek A, Sager R, Lang AB & Schneider H (1997) Protein transport across the in vitro perfused human placenta. Am J Reprod Immunol, 38, 263–271.

Placenta (2000), Vol. 21 Nishida M, Carley WW, Gerritsen ME, Ellingsen O, Kelly RA & Smith TW (1993) Isolation and characterization of human and rat cardiac microvascular endothelial cells. Am J Physiol, 264, H639–H652. Olsen E (1994) Culturing of human umbilical vein and dermal microvascular endothelial cells. In Cell biology. A laboratory handbook, 1 (Ed.) Celis JA, pp. 142–147. Location?: Academic Press. Padykula HA (1988) The human placenta. In Cell and Tissue Biology. A Text Book of Histology (Ed.) Weiss Leon, pp. 899–927. Baltimore, Munich: Urban and Schwarzenberg. Rutten MJ, Hoover RL & Karnovsky MJ (1987) Electrical resistance and macromolecular permeability of brain endothelial monolayer cultures. Brain Res, 425, 301–310. Schnaith E, Beyrau R, Buckner B, Klein RM & Rick W (1994) Optimized determination of ACE activity with hippuryl-Lhistidyl-L-leucine as substrate. Clin Chim Acta, 227, 145–158. Schnitzer JE, Shen C-PJ & Palade GE (1990) Lectin analysis of common glycoproteins detected on the surface of continuos microvascular endothelium in situ and in culture: identification of sialoglycoproteins. Eur J Cell Biol, 52, 241–251. Schu¨tz M & Friedl P (1996) Isolation and cultivation of endothelial cells derived from human placenta. Eur J Cell Biol, 71, 395–401. Schu¨tz M, Teifel M & Friedl P (1997) Establishment of a human endothelial cell line with extended life span after transfection with SV 40 l-antigen. Eur J Cell Biol, 74, 315–320. Shepro D & Morel NML (1993) Pericyte physiology. FASEB J, 7, 1031–1038. Simionescu N & Simionescu M (1976) Galloylglucoses of low molecular weight as mordant in electron microscopy. I. Procedure and evidence for mordanting effect. J Cell Biol, 70, 608–621. Simionescu N & Simionescu M (Eds) (1988) In Endothelial Cell Biology in Health and Disease p. XI. New York: Plenum Press. Simionescu N & Simionescu M (Eds) (1992) In Endothelial Cell Disfunctions p. XI. New York: Plenum Press. Stansby G, Shukla N, Fuller B & Hamilton G (1991) Seeding of human microvascular endothelial cells onto polytetrafluoroethylene graft material. Br J Surg, 78, 1189–1192. Voyta JC, Via DP, Butterfield EC & Zetter BR (1984) Identification and isolation of endothelial cells based on their increased uptake of acetylated LDL. J Cell Biol, 99, 2034–2040. Weibel ER & Palade GE (1964) New cytoplasmic components in arterial endothelial. J Cell Biol, 32, 101–112.