Scand J Haematol (1975) 14, 233-241
Human Fetal Endothelial Cells in Culture TOREHENRIKSEN, STEINA. EVENSEN, M.D., ~ G & ANNE VEFLING
N F A~ L L I N ELGJO, G M.D.
Institute f o r Surgical Research (Chief, E. Amundsen), Medical Department A (Chief, E . Gjone), Rikshospitalet and Department of Pathology (Chief, R . Eker), The Norwegian Radium Hospital, Oslo, Norway
Human endothelial cells were isolated from the umbilical cord vein by collagenase treatment and cultured for periods up to 6 weeks. The cultured cells were identified as endothelium by cell morphology and growth pattern, the presence of WeibelPalade bodies, and their ability to stimulate allogeneic lymphocytes (Hirschberg et al 1974). Cultured fibroblast-like cells derived from the umbilical cord were clearly different in all three respects. Approximately one third of the primary endothelial cultures showed clear evidence of proliferation during the first 3-4 days in culture as judged by cell counting. Replicating ability in a culture was correlated with cell density at the time of seeding. Autoradiography of endothelial cells after exposure to 3H-thymidine showed a 30-fold increase in nuclear labelling from day 1 t o day 3 in culture. The endothelial cells have so far been subcultured three times. K e y words: endothelial cells - culture
Accepted for publication January 25, 1975 Correspondence to: T. Henriksen, Institute for Surgical Research, Rikshospitalet, Oslo, Norway
Vascular endothelium interacts with the formed elements of blood and components of plasma during physiologic haemostasis, thrombosis and transplant rejection. Platelets influence the integrity of endothelium (Gimbrone et a1 1969, Roy & Djerassi 1972), and may also have a nurturing effect on endothelium (Ginsburg & Aster 1972). Serotonin released from platelets is probably taken up by endothelium (Strum & Junod 1972). Endothelial cells contain fibrinolytic Scand J Haernatol (1975) 14
agents (Todd 1959), and synthetize factor VIII antigen and von Willebrand factor (Jaffe et a1 1973b, 1974). Thus, endothelium influences several important haematological reactions, and it is surprising how little is known about its functional characteristics. In the period 1963-1967, however, three important reports appeared. Maruyama (1963) introduced a method whereby large amounts of endothelial cells were isolated from the 16
234
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
umbilical cord vein by enzymatic treatment. A year later an organelle was described in endothelial cells which so far has not been found in other cell types (Weibel & Palade 1964). Finally, in 1967 Spaet & Lejnieks were able to estimate the turnover rate of endothelial cells. These works opened numerous new avenues for investigation, and in 1973 JaEe and associates reported culture of morphologically and immunologically identifiable human endothelial cells for periods up to several months. We have isolated human fetal endothelial cells as described by JafEe et a1 (1973a), and report in this paper on the the growth behavior and morphology of these cells. Parts of the presented data have earlier appeared in abstract form (Henriksen et a1 1974). Several functional aspects of these cells are under investigation, including the cholesterol metabolism and their interaction with allogeneic lymphocytes (Hirschberg et a1 1974). MATERIALS AND METHODS lsolation of endothelial cells. The method described by Jaffe et al (1973a) was followed with slight modifications. Human umbilical cords were obtained under aseptic conditions at vaginal deliveries, in a few instances at Caesarean sections. The cords were transpcrted in sterile vessels containing a cold, magnesium and calciumfree phosphate buffer (PB), supplemented with streptomycin and penicillin. The vein was cannulated in both ends and flushed thoroughly with 100-150 ml PB as soon as possible (usually within 2 h, maximally 3 h). Then the vein lumen was filled with 0.2 % coilagenase (Sigma Chem. Comp., St. Louis, cat. no. C-0130) dissolved in PB until moderately dilated. After 10 rnin incubation at 37OC the vein was flushed with 30 ml PB, and the perfusate was collected in a tube containing 10 ml growth medium. After centrifugation at 200 g for 10 min the pellet was resuspended in 2 4 ml growth medium and transferred to either culture dishes, culture flasks or culture plates with multiple microwells.
Culture of endothelial cells. The cells were cultured in either Medium 199 (GIBCO cat. no. 115 EE) or RPMI 1640 (GIBCO cat. no. 240) with 25 mM Hepes buffer. These media were supplemented with 20 % fetal calf serum, (GIBCO cat. no. 614) L-glutamine (0.3 mg/ml), penicillin (100 U/ml), streptomycin (100 pg/ml), and anti-PPLO (GIBCO cat. no. 522, 0.03 mg/mI). All cultures were incubated at 37O C in a humidified 5 % COzair atmosphere. The medium was changed twice weekly. For subculture, the medium was removed and replaced by 0.2 % trpysin i PB. After gentle tilting, excess trypsin solution was removed, leaving a film over the cell layer. After 2-3 rnin at 37O C the loosened cells were resuspended in fresh growth medium and transferred to new culture vessels. Morphological studies. The culture vessels were checked regularly by phase contrast microscopy. For cytology, cells were fixed in situ with absolute methanol and stained with May-GriinwaldGiemsa stain.
Proliferation studies A. Cell counting. Freshly isolated cells were divided equally on a number of flat-bottomed microwells (Lindbro, Chem. Comp. Inc., New Haven, Conn., cat. no. FB-16-24 TC) and cultured under standard conditions. At intervals the cells were washed twice in situ with 0.9 % saline and harvested by trypsination (0.2 % in PB, 10 rnin at 37O C). The well dispersed cells were counted in a Coulter counter. B. D N A synthesis was studied in parallel to cell counting. After varying number of days in culture the medium in the microwells was replaced by fresh growth medium containing 1 yCi/ml (methyl-3H)-thymidine (Radiochemical Center, Amersham, specific activity 20,000-30,000 mCii mMol). After 24 h incubation the cells were rinsed three times with 0.9 % saline and fixed in absolute methanol for 5 min. The bottoms of the wells were cut off and mounted on glass slides with araldite. They were dipped in photographic emulsion (NTB 2, Eastman Kodak, Roch., N Y) diluted 1:l in distilled water, and developed and stained by standard methods after 14 days.
ENDOTHELIAL CELL CULTURE
Isolation and culture of untbilical cord fibroblast-like cells Small pieces of the umbilical vessel wall and surrounding tissue were fixed to the bottom of culture dishes by the chicken plasma clot technique or by the weight of an over-lying glass piece. Culture conditions and methods of study were identical to those for endothelial cells, except that the fibroblast-like cell cultures had to be incubated for at least 15 min with trypsin in order to release enough cells for subcultures. The first subcultures were possible after 3-4 weeks in culture.
RESULTS
Endothelial cells have been isolated from a total of 150 cords. Repeated flushing and moderate dilatation of the vein during collagenase treatment were important to obtain maximal amounts of endothelial cells with minimal contamination of blood cells. Localized swellings observed during flushing
235
prior to collagenase treatment were interpreted as traumatic injuries and the cord was discarded. Rupture of vein wall due to collagenase digestion of the basement membrane was very rare. Bacterial contamination was not a problem, but twice we had fungus infection in the laboratory. Approximately 0.5-1.0 x lo6 cells were obtained from each 10 cm segment of umbilical vein, giving total numbers per cord in the range 1.0-3.0 x lo6 cells. The cells appeared in aggregates of varying sizes which within 2-6 h had stuck to the bottom of the culture vessel and started to migrate out to form star-shaped colonies. In proliferating cultures monolayers of polygonal cells were formed within a week. The nucleus of endothelial cells was usually oval with 2-3 nucleoli and their size varied within narrow limits (Figure l a ) . During six weeks of culture only small changes in their morphology occurred. Cultured fibro-
Figure 1. a. Mono,layer of endothelial cells after 10 days in culture (left). b. Fibroblast-like cells derived from umbilical cord with irregular shape and size, and growing in overlapping layers (right). Phase contrast. 164
236
T. HENRIKSEN, S. A. EVENSEN. R. F. ELGJO & A. VEFLING
blast-like cells from the umbilical cord were microscopically clearly different from the endothelial cells. The cells and their nuclei varied considerably in shape and size, and they grew in varying patterns and overlapping layers (Figure lb). The two types of cells were also ultrastructurally different. The cultured endothelial cells contained Weibel-Palade bodies and showed specialized cell junctions (Figures 2 and 3). The fibroblast-like cells had neither of these two characteristics. The cells also varied with respect to willingness to be subcultured. While endothelial cells were sluggish and to this date have been subcultured only
three times before they failed to make monolayer, the fetal fibroblast-like cells grew rapidly and could be subcultured repeatedly. They attached firmly to the culture vessel and trypsin treatment for 20 min at 37O C only loosened approximately half of the cells. In contrast, endothelial cells lost their attachment during short periods of enzyme treatment (2-3 min). Contamination of the endothelial cell culture by other cells was rare. In a few instances a second cell type was distinguishable which was larger and more spindleshaped. These cells did not show the tendency to overgrow the endothelial cells
Figure 2. Transmission electron micrograph of cultured endothelial cells. Details from 3 neighbouring cells. The arrows point t o specialized areas of the cell junctions where the intercellular lumen is narrowed, and the density of the cytoplasm is increased. The cytoplasma contains vesicles and vacuoles of various sizes, mitochondriae (M), single microfibrils (F), bundles of fibres (FB), rough endoplasmic reticulum (RER), and clusters of free ribosomes. An autophagic vacuole (AV) and a Weibel-Palade body in cross section are also shown (x 30,000).
237
ENDOTHELIAL CELL CULTURE
and often appeared to be left in the culture vessel during the process of subculturing. Proliferation studies
Freshly isolated cells could be distributed in microwells with less than 10 % variance between wells. Using the cell counts 24 h after seeding as base-line, it could be shown that cell counts increased in 35 % of primary cultures during the next 3-4 days, showed no significant change in 36 % and decreased in 29 %. Among the latter, 20 % showed an increase in cell numbers in the next 4 days. In cultures with increasing cell counts large variations were found, with 28 and 169 % increase as the extremes. Fig-
ure 4 shows how cell density influenced proliferation. High cell density obviously inhibited proliferation; in cultures with more than 1000 cells per mm2 of culture vessel surface, significant increase in cell counts was never observed. In contrast, more than half of the cultures with less than 750 cells per mm2 showed evidence of proliferation, with an apparent optimum in the range 500-750 cells per mm2. Cell counts were also followed from day 3 to day 6-7 (not shown in the Figure). Again, cultures with low density at day 3 (less than 500 cells per mm2) showed the best ability to proliferate. In primary cultures with rapid proliferation during the first three days, cell counts often decreased in the following days. Separate experiments comparing proliferation of cells from the same donor in medium 199 and RPMI 1640, respectively, (with identical supplementation of serum,
t
......
C"L1"RES
n CELLS
Figure 3. Transmission electron micrograph of endothelial cells cultured fosr 2 weeks. Several WP bodies are shown both in longitudinal and cross sections.
PROLIFERATING
PER M M ~
Figure 4. Effect of cell density on the proliferation of endothelial cells in culture. Each column shows the number of cultures studied within each density group in the interval 24 h to 72-96 h after seeding. Hatched areas of the columns represent the number of cultures which showed significant increase in cell counts.
238
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
TABLE I Incidence o f labelled nuclei in endothelial cell cultures grown for 24, 72 or I 9 2 h before being exposed to 3H-thymidine
24 h
c 1 1 1 -
I
72 h
192 h
Labelled nuclei/ 1 Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ total nuclei total nuclei total nuclei 104 nuclei 104 nuclei 1 lo4 nuclei counted counted counted 1
I
I
1 2 3
313246 0/2056 6/1950 12/1854
4
9 0 31 65
212/2074 113/2126 12712024 233/2155
26
Mean
nutrients and antibiotics) did not disclose any difference during a period of 14 days. For comparison, the proliferation of fetal fibroblast-like cells cultured for 5-8 weeks is shown in Figure 5. These studies were performed as described for endothelial cells.
18 2o 16
I
N
0 ’
14
“E 12 L .
10
Y 1 W
8
0
6 L
2
I 2 4 6 8 1012 DAYS AFTER SUBCULTURE Figure 5. Proliferation of fetal fibroblast-like cells kept in culture for 5-8 weeks and subcultured 2-5 times. The curve represents the mean of 5 individual series of cell counts.
1022 531 627 1081 815
19/2016 73/1947 140/2013 118/2023
94 375 695 583 437
In contrast to what was observed for endothelial cells, no inhibition of proliferation was observed at higher cell densities. Cultures with initial levels of above 1000 cells per mm2 showed the same pattern (not shown in the Figure). Older fibroblast-like cells subcultured a number of times (more than 11 weeks old) showed an overall slower growth rate, especially at higher cell densites. Human skin fibroblast-like cells rLsembled freshly isolated fetal fibroblasts in culture, but appeared to proliferate slightly better in medium RPMI 1640 than in medium 199. Autoradiographic studies supported the results of cell counting (Table I). While very few labelled nuclei were found in 24 h old cultures, the number of endothelial cells engaged in DNA synthesis had increased 30fold 72 h after seeding. After 8 days in primary culture a number of cells still showed thymidine uptake, but distinctly fewer than at 72 h. The number of labelled nuclei was highest in areas of the cultures where the cell density was relatively low. The average cell density in cultures included in the autoradiographical studes was within the range of optimal cell density for proliferation as judged by cell counting (Figure 4).
ENDOTHELIAL CELL CULTURE
239
TABLE 11 Characteristics of umbilical vein endothelial cells Umbilical vein endothelial cells
Umbilical cord fibroblast-like cells
Cell morphology
polygonal with specialized cell junctions and of uniform size
spindle-shaped, variable cell and nuclear size
Growth pattern
monolayer
overlapping layers
Attachment to culture vessel
loose, cells easily released by trypsin
firm, cells slowly released by trypsin
Weibel-Palade bodies
present
not present
Stimulation of allogeneic lymphocytes (Hirschberg et a1 1974)
yes
no
DISCUSSION
Culture of endothelial cells has been attempted by several investigators (Maruyama 1963, Pomerat & Slick 1963, Fryer et a1 1966, Wechezak & Mansfield 1973), but it was not until 1973 that Jaffe and associates reported long-time culture of isolated cells from umbilical veins which could be identified as endothelium by morphological and immunological criteria. In this article we confirm and extend their observations. A summary of the criteria we have used for identification of the isolated endothelial cells is given in Table 11. For comparison we also cultured fibroblast-like cells from the umbilical cord. The endothelial cells had a typical morphology and growth pattern, but that alone did not prove their identity. Probably the most decisive characteristic was the presence of Weibel-Palade bodies (Figure 3). This rod-shaped tubular organelle of unknown function, described by Weibel & Palade in 1964, is by now generally accepted as specific of endothelial cells (Jaffe et a1 1973a, McDonald et a1 1973, Lewis et a1 1973, Gimbrone et a1 1974, Henriksen et a1 3974). A detailed
report on their occurrence and morphology, and the ultrastructure of fetal endothelium in general will be reported elsewhere. The ability of endothelial cells to stimulate allogeneic lymphocytes has also proved useful in the differentiation of fetal endothelial cells from cord fibroblast-like cells (Hirschberg et a1 1974).Vetto & Burger (2972) first showed that canine lymphocytes were stimulated by allogeneic endothelial cells. Preliminary observations indicate that endothelial cells, suspended in platelet-poor plasma markedly accelerate the degradation of ADP to adenosine, while human skin fibroblasts do not (Henriksen et a1 1974). Other suggested differentiating characteristics as compared to human skin fibroblasts are the presence of smooth muscle actomyosin and blood group antigens (Jaffe et a1 1973a) and the synthesis of antihaemophilic factor antigen (Jaffe et a1 1973b). Contamination of the cultures by cells morphologically distinct from endothelial cells was rare. We do not know whether they represent fibroblasts or smooth muscle cells. Similar contaminants were reported by
240
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
Gimbrone et a1 (1974) who convincingly showed them to be smooth mucsle cells. Overgrowth of endothelial cultures by smooth muscle cells occurs if the cord is deliberately damaged by repeatedly clamping before collagenase treatment (Jaffe et a1 1973a). After Spaet & Lejnieks original observation in 1967 several studies on the turnover rate of the endothelial cells in situ after exposure to various vasoactive drugs and trauma have appeared (Gaynor 1971, Sade et a1 1972, Tannock & Hayashi 1972, Evensen et a1 1973, Evensen & Shepro 1974). In this paper the increasing cell counts and the demonstration of accelerated DNA synthesis prove the replicating ability of the cultured cells, but the proliferative activity was variable. Approximately one third of the primary cultures showed definitely increased cell counts within the first 3-4 days. Comparable data are not at hand, but Gimbrone et a1 (1974) stated that one third of their primary cultures did not reach confluent densities. Cell doubling times vary considerably in the available reports; from 12-24 h (Mc Donald et a1 1973) to 92 h (Jaf€e et a1 1973a). Our proliferating primary cultures had a cell doubling time of approximately 72 h. Low cell density in the primary culture was correlated with ability to proliferate, while further replication did not appear to take place at cell densities above 1000 per nun2. Similar figures are given by Gimbrone et a1 (1974). So far we have not been able to subculture endothelial cells more than three times before they failed to reach confluency. Other investigators have reported numerous passages (McDonald et a1 1973, Jaffe et a1 1973a, Gimbrone et a1 1974). The reason for this discrepancy is not yet understood.
ACKNOWLEDGEMENTS We are grateful for the kind cooperation of Professor K. Bj@ro and Dr. P. Fylling and the staffs at Kvinneklinikken, Rikshospitalet and Oslo Kommunale Kvinneklinikk, Oslo, in obtaining umbilical cords. We also thank Dr. E. Jaffe and Dr. R. Nachman for helpful suggestions with the isolation procedure. This work was supported in part by the Norwegian Research Council for Science and the Humanities, the Norwegian Council on Cardiovascular Diseases, and legacies from the University of Oslo.
REFERENCES Evensen S A, Elgjo R F & Shepro D (1973) Platelets, endothelium and the triggering mechanism of intravascular coagulation. Thromb Diath Haemorrh, Suppl. 54, 207-10. Evensen S A & Shepro D (1974) DNA synthesis in rat aortic endothelium: effect of bacterial endotoxin and trauma. Microvasc Res 8, 90-96. Fryer D G, Birnbaum G & Luttrell C N (1966) Human endothelium in cell culture. J Atheroscler Res 6 , 151-63. Gaynor E (1971) Increased mitotic activity in rabbit endothelium after endotoxin. Lab Znvest 24, 318-20. Girnbrone M A Jr, Aster R H, Cotran R S, Corkery J, Jandl J H & Folkman J (1969) Preservation of vascular integrity in organs perfused in vitro with a platelet-rich medium. Nature 222, 33-36. Gimbrone M A, Cotran R S & Folkman J (1974) Human vascular endothelial cells in culture. Growth and DNA synthesis. J Cell Biol 60, 673-84. Ginsburg A D & Aster R H (1972) The nurturing effect of platelets on vascular endothelium. Thromb Diath Haemorrh, Suppl. 51, 143-49. Henriksen T, Evensen S A, Elgjo R F, Hirschberg H, Holmsen I & Vefling A (1974) Properties of human endothelial cells in culture. Proc X V Congr Znternat Soc Hematol, Jerusalem 1974, p 167. Hirschberg H, Evensen S A, Henriksen T & Thorsby E (1974) Stimulation of human lymphocytes by allogeneic endothelial cells in vitro. Tissue Antigens 4, 257-61.
ENDOTHELIAL CELL CULTURE Jaffe E A, Nachman R L, Becker C G & Minick C R (1973a) Culture of human endothelial cells derived from umbilical veins. J Clin Invest 52, 2745-56. Jaffe E A, Hoyer L W & Nachman R L (1973b) Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J Clin Invest 52, 2757-64. Jaffe E A, Hoyer L W & Nachman R L (1974) Synthesis of von Willebrand factor by cultured human endothelial cells. Proc Natl Acad Sci USA 71, 1906-09. Lewis L J, Hoak J C, Maca R D & Fry G L (1973) Replication of human endothelial cells in culture. Science 181, 453-54. Maruyama Y (1963) The human endothelial cell in tissue culture. Z Zellforsch Mikrosk Anat 60, 69-79. McDonald R I, Shepro D, Rosenthal M & Booyse F M (1973) Properties of cultured endothelial cells. Ser Haematol VI, 4, 469-78. Pomerat C M & Slick W C (1963) Isolation and growth of endothelial cells in tissue culture. Nature 198, 859-61. Roy A J & Djerassi I (1972) Effects of platelet transfusions: plug formation and maintenance
24 1
of vascular integrity. Proc SOC Exp Biol Med 139, 137-42. Sade R M, Folkman J & Cotran R S (1972) DNA synthesis in endothelium of aortic segments in vitro. Exp Cell Res 74, 297-306. Spaet T H & Lejnieks I (1967) Mitotic activity of rabbit blood vessels. Proc SOC Exp Biol Med 125, 1197-1201. Strum J M & Junod A F (1972) Radioautographic demonstration of 5-hydro~ytryptamine-~H uptake by pulmonary endothelial cells. J Cell Biol 54, 456-67. Tannock I F & Hayashi S (1972) The proliferation of capillary endothelial cells. Cancer Res 32, 77-82. Todd A S (1959) The histological localization of fibrinolysis activator. J Pathol 78, 281-83. Vetto R M & Burger D R (1972) Endothelial cell stimulation of allogeneic lymphocytes. Transplantation 14, 652-54. Wechezak A R & Mansfield P B (1973) Isolation and growth characteristics of cell lines from bovine venous endothelium. In Vitro 9, 39-45. Weibel E R & Palade G E (1964) New cytoplasmic components in arterial endothelia. J Cell Biol 23, 101-12.
Human Fetal Endothelial Cells in Culture TOREHENRIKSEN, STEINA. EVENSEN, M.D., ~ G & ANNE VEFLING
N F A~ L L I N ELGJO, G M.D.
Institute f o r Surgical Research (Chief, E. Amundsen), Medical Department A (Chief, E . Gjone), Rikshospitalet and Department of Pathology (Chief, R . Eker), The Norwegian Radium Hospital, Oslo, Norway
Human endothelial cells were isolated from the umbilical cord vein by collagenase treatment and cultured for periods up to 6 weeks. The cultured cells were identified as endothelium by cell morphology and growth pattern, the presence of WeibelPalade bodies, and their ability to stimulate allogeneic lymphocytes (Hirschberg et al 1974). Cultured fibroblast-like cells derived from the umbilical cord were clearly different in all three respects. Approximately one third of the primary endothelial cultures showed clear evidence of proliferation during the first 3-4 days in culture as judged by cell counting. Replicating ability in a culture was correlated with cell density at the time of seeding. Autoradiography of endothelial cells after exposure to 3H-thymidine showed a 30-fold increase in nuclear labelling from day 1 t o day 3 in culture. The endothelial cells have so far been subcultured three times. K e y words: endothelial cells - culture
Accepted for publication January 25, 1975 Correspondence to: T. Henriksen, Institute for Surgical Research, Rikshospitalet, Oslo, Norway
Vascular endothelium interacts with the formed elements of blood and components of plasma during physiologic haemostasis, thrombosis and transplant rejection. Platelets influence the integrity of endothelium (Gimbrone et a1 1969, Roy & Djerassi 1972), and may also have a nurturing effect on endothelium (Ginsburg & Aster 1972). Serotonin released from platelets is probably taken up by endothelium (Strum & Junod 1972). Endothelial cells contain fibrinolytic Scand J Haernatol (1975) 14
agents (Todd 1959), and synthetize factor VIII antigen and von Willebrand factor (Jaffe et a1 1973b, 1974). Thus, endothelium influences several important haematological reactions, and it is surprising how little is known about its functional characteristics. In the period 1963-1967, however, three important reports appeared. Maruyama (1963) introduced a method whereby large amounts of endothelial cells were isolated from the 16
234
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
umbilical cord vein by enzymatic treatment. A year later an organelle was described in endothelial cells which so far has not been found in other cell types (Weibel & Palade 1964). Finally, in 1967 Spaet & Lejnieks were able to estimate the turnover rate of endothelial cells. These works opened numerous new avenues for investigation, and in 1973 JaEe and associates reported culture of morphologically and immunologically identifiable human endothelial cells for periods up to several months. We have isolated human fetal endothelial cells as described by JafEe et a1 (1973a), and report in this paper on the the growth behavior and morphology of these cells. Parts of the presented data have earlier appeared in abstract form (Henriksen et a1 1974). Several functional aspects of these cells are under investigation, including the cholesterol metabolism and their interaction with allogeneic lymphocytes (Hirschberg et a1 1974). MATERIALS AND METHODS lsolation of endothelial cells. The method described by Jaffe et al (1973a) was followed with slight modifications. Human umbilical cords were obtained under aseptic conditions at vaginal deliveries, in a few instances at Caesarean sections. The cords were transpcrted in sterile vessels containing a cold, magnesium and calciumfree phosphate buffer (PB), supplemented with streptomycin and penicillin. The vein was cannulated in both ends and flushed thoroughly with 100-150 ml PB as soon as possible (usually within 2 h, maximally 3 h). Then the vein lumen was filled with 0.2 % coilagenase (Sigma Chem. Comp., St. Louis, cat. no. C-0130) dissolved in PB until moderately dilated. After 10 rnin incubation at 37OC the vein was flushed with 30 ml PB, and the perfusate was collected in a tube containing 10 ml growth medium. After centrifugation at 200 g for 10 min the pellet was resuspended in 2 4 ml growth medium and transferred to either culture dishes, culture flasks or culture plates with multiple microwells.
Culture of endothelial cells. The cells were cultured in either Medium 199 (GIBCO cat. no. 115 EE) or RPMI 1640 (GIBCO cat. no. 240) with 25 mM Hepes buffer. These media were supplemented with 20 % fetal calf serum, (GIBCO cat. no. 614) L-glutamine (0.3 mg/ml), penicillin (100 U/ml), streptomycin (100 pg/ml), and anti-PPLO (GIBCO cat. no. 522, 0.03 mg/mI). All cultures were incubated at 37O C in a humidified 5 % COzair atmosphere. The medium was changed twice weekly. For subculture, the medium was removed and replaced by 0.2 % trpysin i PB. After gentle tilting, excess trypsin solution was removed, leaving a film over the cell layer. After 2-3 rnin at 37O C the loosened cells were resuspended in fresh growth medium and transferred to new culture vessels. Morphological studies. The culture vessels were checked regularly by phase contrast microscopy. For cytology, cells were fixed in situ with absolute methanol and stained with May-GriinwaldGiemsa stain.
Proliferation studies A. Cell counting. Freshly isolated cells were divided equally on a number of flat-bottomed microwells (Lindbro, Chem. Comp. Inc., New Haven, Conn., cat. no. FB-16-24 TC) and cultured under standard conditions. At intervals the cells were washed twice in situ with 0.9 % saline and harvested by trypsination (0.2 % in PB, 10 rnin at 37O C). The well dispersed cells were counted in a Coulter counter. B. D N A synthesis was studied in parallel to cell counting. After varying number of days in culture the medium in the microwells was replaced by fresh growth medium containing 1 yCi/ml (methyl-3H)-thymidine (Radiochemical Center, Amersham, specific activity 20,000-30,000 mCii mMol). After 24 h incubation the cells were rinsed three times with 0.9 % saline and fixed in absolute methanol for 5 min. The bottoms of the wells were cut off and mounted on glass slides with araldite. They were dipped in photographic emulsion (NTB 2, Eastman Kodak, Roch., N Y) diluted 1:l in distilled water, and developed and stained by standard methods after 14 days.
ENDOTHELIAL CELL CULTURE
Isolation and culture of untbilical cord fibroblast-like cells Small pieces of the umbilical vessel wall and surrounding tissue were fixed to the bottom of culture dishes by the chicken plasma clot technique or by the weight of an over-lying glass piece. Culture conditions and methods of study were identical to those for endothelial cells, except that the fibroblast-like cell cultures had to be incubated for at least 15 min with trypsin in order to release enough cells for subcultures. The first subcultures were possible after 3-4 weeks in culture.
RESULTS
Endothelial cells have been isolated from a total of 150 cords. Repeated flushing and moderate dilatation of the vein during collagenase treatment were important to obtain maximal amounts of endothelial cells with minimal contamination of blood cells. Localized swellings observed during flushing
235
prior to collagenase treatment were interpreted as traumatic injuries and the cord was discarded. Rupture of vein wall due to collagenase digestion of the basement membrane was very rare. Bacterial contamination was not a problem, but twice we had fungus infection in the laboratory. Approximately 0.5-1.0 x lo6 cells were obtained from each 10 cm segment of umbilical vein, giving total numbers per cord in the range 1.0-3.0 x lo6 cells. The cells appeared in aggregates of varying sizes which within 2-6 h had stuck to the bottom of the culture vessel and started to migrate out to form star-shaped colonies. In proliferating cultures monolayers of polygonal cells were formed within a week. The nucleus of endothelial cells was usually oval with 2-3 nucleoli and their size varied within narrow limits (Figure l a ) . During six weeks of culture only small changes in their morphology occurred. Cultured fibro-
Figure 1. a. Mono,layer of endothelial cells after 10 days in culture (left). b. Fibroblast-like cells derived from umbilical cord with irregular shape and size, and growing in overlapping layers (right). Phase contrast. 164
236
T. HENRIKSEN, S. A. EVENSEN. R. F. ELGJO & A. VEFLING
blast-like cells from the umbilical cord were microscopically clearly different from the endothelial cells. The cells and their nuclei varied considerably in shape and size, and they grew in varying patterns and overlapping layers (Figure lb). The two types of cells were also ultrastructurally different. The cultured endothelial cells contained Weibel-Palade bodies and showed specialized cell junctions (Figures 2 and 3). The fibroblast-like cells had neither of these two characteristics. The cells also varied with respect to willingness to be subcultured. While endothelial cells were sluggish and to this date have been subcultured only
three times before they failed to make monolayer, the fetal fibroblast-like cells grew rapidly and could be subcultured repeatedly. They attached firmly to the culture vessel and trypsin treatment for 20 min at 37O C only loosened approximately half of the cells. In contrast, endothelial cells lost their attachment during short periods of enzyme treatment (2-3 min). Contamination of the endothelial cell culture by other cells was rare. In a few instances a second cell type was distinguishable which was larger and more spindleshaped. These cells did not show the tendency to overgrow the endothelial cells
Figure 2. Transmission electron micrograph of cultured endothelial cells. Details from 3 neighbouring cells. The arrows point t o specialized areas of the cell junctions where the intercellular lumen is narrowed, and the density of the cytoplasm is increased. The cytoplasma contains vesicles and vacuoles of various sizes, mitochondriae (M), single microfibrils (F), bundles of fibres (FB), rough endoplasmic reticulum (RER), and clusters of free ribosomes. An autophagic vacuole (AV) and a Weibel-Palade body in cross section are also shown (x 30,000).
237
ENDOTHELIAL CELL CULTURE
and often appeared to be left in the culture vessel during the process of subculturing. Proliferation studies
Freshly isolated cells could be distributed in microwells with less than 10 % variance between wells. Using the cell counts 24 h after seeding as base-line, it could be shown that cell counts increased in 35 % of primary cultures during the next 3-4 days, showed no significant change in 36 % and decreased in 29 %. Among the latter, 20 % showed an increase in cell numbers in the next 4 days. In cultures with increasing cell counts large variations were found, with 28 and 169 % increase as the extremes. Fig-
ure 4 shows how cell density influenced proliferation. High cell density obviously inhibited proliferation; in cultures with more than 1000 cells per mm2 of culture vessel surface, significant increase in cell counts was never observed. In contrast, more than half of the cultures with less than 750 cells per mm2 showed evidence of proliferation, with an apparent optimum in the range 500-750 cells per mm2. Cell counts were also followed from day 3 to day 6-7 (not shown in the Figure). Again, cultures with low density at day 3 (less than 500 cells per mm2) showed the best ability to proliferate. In primary cultures with rapid proliferation during the first three days, cell counts often decreased in the following days. Separate experiments comparing proliferation of cells from the same donor in medium 199 and RPMI 1640, respectively, (with identical supplementation of serum,
t
......
C"L1"RES
n CELLS
Figure 3. Transmission electron micrograph of endothelial cells cultured fosr 2 weeks. Several WP bodies are shown both in longitudinal and cross sections.
PROLIFERATING
PER M M ~
Figure 4. Effect of cell density on the proliferation of endothelial cells in culture. Each column shows the number of cultures studied within each density group in the interval 24 h to 72-96 h after seeding. Hatched areas of the columns represent the number of cultures which showed significant increase in cell counts.
238
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
TABLE I Incidence o f labelled nuclei in endothelial cell cultures grown for 24, 72 or I 9 2 h before being exposed to 3H-thymidine
24 h
c 1 1 1 -
I
72 h
192 h
Labelled nuclei/ 1 Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ Labelled nuclei/ total nuclei total nuclei total nuclei 104 nuclei 104 nuclei 1 lo4 nuclei counted counted counted 1
I
I
1 2 3
313246 0/2056 6/1950 12/1854
4
9 0 31 65
212/2074 113/2126 12712024 233/2155
26
Mean
nutrients and antibiotics) did not disclose any difference during a period of 14 days. For comparison, the proliferation of fetal fibroblast-like cells cultured for 5-8 weeks is shown in Figure 5. These studies were performed as described for endothelial cells.
18 2o 16
I
N
0 ’
14
“E 12 L .
10
Y 1 W
8
0
6 L
2
I 2 4 6 8 1012 DAYS AFTER SUBCULTURE Figure 5. Proliferation of fetal fibroblast-like cells kept in culture for 5-8 weeks and subcultured 2-5 times. The curve represents the mean of 5 individual series of cell counts.
1022 531 627 1081 815
19/2016 73/1947 140/2013 118/2023
94 375 695 583 437
In contrast to what was observed for endothelial cells, no inhibition of proliferation was observed at higher cell densities. Cultures with initial levels of above 1000 cells per mm2 showed the same pattern (not shown in the Figure). Older fibroblast-like cells subcultured a number of times (more than 11 weeks old) showed an overall slower growth rate, especially at higher cell densites. Human skin fibroblast-like cells rLsembled freshly isolated fetal fibroblasts in culture, but appeared to proliferate slightly better in medium RPMI 1640 than in medium 199. Autoradiographic studies supported the results of cell counting (Table I). While very few labelled nuclei were found in 24 h old cultures, the number of endothelial cells engaged in DNA synthesis had increased 30fold 72 h after seeding. After 8 days in primary culture a number of cells still showed thymidine uptake, but distinctly fewer than at 72 h. The number of labelled nuclei was highest in areas of the cultures where the cell density was relatively low. The average cell density in cultures included in the autoradiographical studes was within the range of optimal cell density for proliferation as judged by cell counting (Figure 4).
ENDOTHELIAL CELL CULTURE
239
TABLE 11 Characteristics of umbilical vein endothelial cells Umbilical vein endothelial cells
Umbilical cord fibroblast-like cells
Cell morphology
polygonal with specialized cell junctions and of uniform size
spindle-shaped, variable cell and nuclear size
Growth pattern
monolayer
overlapping layers
Attachment to culture vessel
loose, cells easily released by trypsin
firm, cells slowly released by trypsin
Weibel-Palade bodies
present
not present
Stimulation of allogeneic lymphocytes (Hirschberg et a1 1974)
yes
no
DISCUSSION
Culture of endothelial cells has been attempted by several investigators (Maruyama 1963, Pomerat & Slick 1963, Fryer et a1 1966, Wechezak & Mansfield 1973), but it was not until 1973 that Jaffe and associates reported long-time culture of isolated cells from umbilical veins which could be identified as endothelium by morphological and immunological criteria. In this article we confirm and extend their observations. A summary of the criteria we have used for identification of the isolated endothelial cells is given in Table 11. For comparison we also cultured fibroblast-like cells from the umbilical cord. The endothelial cells had a typical morphology and growth pattern, but that alone did not prove their identity. Probably the most decisive characteristic was the presence of Weibel-Palade bodies (Figure 3). This rod-shaped tubular organelle of unknown function, described by Weibel & Palade in 1964, is by now generally accepted as specific of endothelial cells (Jaffe et a1 1973a, McDonald et a1 1973, Lewis et a1 1973, Gimbrone et a1 1974, Henriksen et a1 3974). A detailed
report on their occurrence and morphology, and the ultrastructure of fetal endothelium in general will be reported elsewhere. The ability of endothelial cells to stimulate allogeneic lymphocytes has also proved useful in the differentiation of fetal endothelial cells from cord fibroblast-like cells (Hirschberg et a1 1974).Vetto & Burger (2972) first showed that canine lymphocytes were stimulated by allogeneic endothelial cells. Preliminary observations indicate that endothelial cells, suspended in platelet-poor plasma markedly accelerate the degradation of ADP to adenosine, while human skin fibroblasts do not (Henriksen et a1 1974). Other suggested differentiating characteristics as compared to human skin fibroblasts are the presence of smooth muscle actomyosin and blood group antigens (Jaffe et a1 1973a) and the synthesis of antihaemophilic factor antigen (Jaffe et a1 1973b). Contamination of the cultures by cells morphologically distinct from endothelial cells was rare. We do not know whether they represent fibroblasts or smooth muscle cells. Similar contaminants were reported by
240
T. HENRIKSEN, S. A. EVENSEN, R. F. ELGJO & A. VEFLING
Gimbrone et a1 (1974) who convincingly showed them to be smooth mucsle cells. Overgrowth of endothelial cultures by smooth muscle cells occurs if the cord is deliberately damaged by repeatedly clamping before collagenase treatment (Jaffe et a1 1973a). After Spaet & Lejnieks original observation in 1967 several studies on the turnover rate of the endothelial cells in situ after exposure to various vasoactive drugs and trauma have appeared (Gaynor 1971, Sade et a1 1972, Tannock & Hayashi 1972, Evensen et a1 1973, Evensen & Shepro 1974). In this paper the increasing cell counts and the demonstration of accelerated DNA synthesis prove the replicating ability of the cultured cells, but the proliferative activity was variable. Approximately one third of the primary cultures showed definitely increased cell counts within the first 3-4 days. Comparable data are not at hand, but Gimbrone et a1 (1974) stated that one third of their primary cultures did not reach confluent densities. Cell doubling times vary considerably in the available reports; from 12-24 h (Mc Donald et a1 1973) to 92 h (Jaf€e et a1 1973a). Our proliferating primary cultures had a cell doubling time of approximately 72 h. Low cell density in the primary culture was correlated with ability to proliferate, while further replication did not appear to take place at cell densities above 1000 per nun2. Similar figures are given by Gimbrone et a1 (1974). So far we have not been able to subculture endothelial cells more than three times before they failed to reach confluency. Other investigators have reported numerous passages (McDonald et a1 1973, Jaffe et a1 1973a, Gimbrone et a1 1974). The reason for this discrepancy is not yet understood.
ACKNOWLEDGEMENTS We are grateful for the kind cooperation of Professor K. Bj@ro and Dr. P. Fylling and the staffs at Kvinneklinikken, Rikshospitalet and Oslo Kommunale Kvinneklinikk, Oslo, in obtaining umbilical cords. We also thank Dr. E. Jaffe and Dr. R. Nachman for helpful suggestions with the isolation procedure. This work was supported in part by the Norwegian Research Council for Science and the Humanities, the Norwegian Council on Cardiovascular Diseases, and legacies from the University of Oslo.
REFERENCES Evensen S A, Elgjo R F & Shepro D (1973) Platelets, endothelium and the triggering mechanism of intravascular coagulation. Thromb Diath Haemorrh, Suppl. 54, 207-10. Evensen S A & Shepro D (1974) DNA synthesis in rat aortic endothelium: effect of bacterial endotoxin and trauma. Microvasc Res 8, 90-96. Fryer D G, Birnbaum G & Luttrell C N (1966) Human endothelium in cell culture. J Atheroscler Res 6 , 151-63. Gaynor E (1971) Increased mitotic activity in rabbit endothelium after endotoxin. Lab Znvest 24, 318-20. Girnbrone M A Jr, Aster R H, Cotran R S, Corkery J, Jandl J H & Folkman J (1969) Preservation of vascular integrity in organs perfused in vitro with a platelet-rich medium. Nature 222, 33-36. Gimbrone M A, Cotran R S & Folkman J (1974) Human vascular endothelial cells in culture. Growth and DNA synthesis. J Cell Biol 60, 673-84. Ginsburg A D & Aster R H (1972) The nurturing effect of platelets on vascular endothelium. Thromb Diath Haemorrh, Suppl. 51, 143-49. Henriksen T, Evensen S A, Elgjo R F, Hirschberg H, Holmsen I & Vefling A (1974) Properties of human endothelial cells in culture. Proc X V Congr Znternat Soc Hematol, Jerusalem 1974, p 167. Hirschberg H, Evensen S A, Henriksen T & Thorsby E (1974) Stimulation of human lymphocytes by allogeneic endothelial cells in vitro. Tissue Antigens 4, 257-61.
ENDOTHELIAL CELL CULTURE Jaffe E A, Nachman R L, Becker C G & Minick C R (1973a) Culture of human endothelial cells derived from umbilical veins. J Clin Invest 52, 2745-56. Jaffe E A, Hoyer L W & Nachman R L (1973b) Synthesis of antihemophilic factor antigen by cultured human endothelial cells. J Clin Invest 52, 2757-64. Jaffe E A, Hoyer L W & Nachman R L (1974) Synthesis of von Willebrand factor by cultured human endothelial cells. Proc Natl Acad Sci USA 71, 1906-09. Lewis L J, Hoak J C, Maca R D & Fry G L (1973) Replication of human endothelial cells in culture. Science 181, 453-54. Maruyama Y (1963) The human endothelial cell in tissue culture. Z Zellforsch Mikrosk Anat 60, 69-79. McDonald R I, Shepro D, Rosenthal M & Booyse F M (1973) Properties of cultured endothelial cells. Ser Haematol VI, 4, 469-78. Pomerat C M & Slick W C (1963) Isolation and growth of endothelial cells in tissue culture. Nature 198, 859-61. Roy A J & Djerassi I (1972) Effects of platelet transfusions: plug formation and maintenance
24 1
of vascular integrity. Proc SOC Exp Biol Med 139, 137-42. Sade R M, Folkman J & Cotran R S (1972) DNA synthesis in endothelium of aortic segments in vitro. Exp Cell Res 74, 297-306. Spaet T H & Lejnieks I (1967) Mitotic activity of rabbit blood vessels. Proc SOC Exp Biol Med 125, 1197-1201. Strum J M & Junod A F (1972) Radioautographic demonstration of 5-hydro~ytryptamine-~H uptake by pulmonary endothelial cells. J Cell Biol 54, 456-67. Tannock I F & Hayashi S (1972) The proliferation of capillary endothelial cells. Cancer Res 32, 77-82. Todd A S (1959) The histological localization of fibrinolysis activator. J Pathol 78, 281-83. Vetto R M & Burger D R (1972) Endothelial cell stimulation of allogeneic lymphocytes. Transplantation 14, 652-54. Wechezak A R & Mansfield P B (1973) Isolation and growth characteristics of cell lines from bovine venous endothelium. In Vitro 9, 39-45. Weibel E R & Palade G E (1964) New cytoplasmic components in arterial endothelia. J Cell Biol 23, 101-12.
