SYNTHESIS OF FACTOR VIII ANTIGEN BY CULTURED HUMAN ENDOTHELIAL CELLS * Eric A. Jaffe Division of Hematology Department of Medicine The New York Hospital-Cornell Medical Center New York, New York 10021
Factor VIII (antihemophilic factor, AHF) is a plasma glycoprotein that in the purified state has a molecular weight of 1.12 x 1OS.l The partially purified protein has three distinct properties: (a) factor VIII antigen (VIIIAGX), an antigen identified in precipitin assays by heterologous antibodies and decreased in plasma of patients with von Willebrand’s disease but normal in hemophilic plasma; (b) factor VIII clot promoting function (VIIIAHF), an activity that corrects the plasma coagulation abnormalities of patients with classic hemophilia (hemophilia A ) ; and (c) von Willebrand factor (VIIIAHF), an activity deficient in von Willebrand‘s disease, which can be identified by an abnormality of the bleeding time or by in vitro assays of platelet function (retention of platelets in glass bead columns o r ristocetin-induced aggregation of washed normal human platelets). The term “factor VIII” identifies a protein(s) present in normal human plasma which is (are) responsible for these three activities. This operational nomenclature has recently been suggested by Weiss and coworkers.2 The site of productions of factor VIII is unknown. The l i ~ e r , ~ - ~ spleen,3* 6 , kidney,x reticuloendothelial cell: and lymphocytes lo have all been suggested as sites of synthesis. Recent immunofluorescence studies performed on a variety of human tissue sections, using a monospecific rabbit antihuman factor VIII, have demonstrated that endothelial cells contain VIIIAGs.11-13The development in our laboratory of methods for culturing and identifying human endothelial cells l 4 3 1s has made it possible to examine the role of the endothelial cell in factor VIII synthesis. Our results show that cultured human endothelial cells synthesize and release VIIIAGs into their culture medium. 4v
RESULTS Culture of Human Endothelial Cells
Endothelial cells were obtained from human umbilical cord veins by digestion of the vein’s interior with 0.2% collagenase for 15 minutes at 37” C. The collagenase solution was drained from the vein, which was then rinsed with cord buffer (0.14 M NaCI, 0.004 M KCl, 0.001 M phosphate buffer, pH 7.4, 0.01 1 M glucose). The cells present in these solutions were sedimented by centrifugation at 250 g, washed once with Medium 199 containing 20% :ii Supported by grants from the National Institutes of Health (HL-14810)and The Arnold R. Krakower Hematology Foundation.
62
Jaffe: Synthesis of Factor VIII Antigen
63
fetal calf serum, and cultured in this medium at 37" under 5 % C02.15 Smoothmuscle cells and fibroblasts were also cultured as previously described.15 After one week in culture, endothelial cells grew in confluent monolayers without a definable whorling pattern. The cells were large (30 x 50 p ) , polygonal, and closely apposed with an oval, centrally located nucleus and indistinct cell borders (FIGURE1). The cells formed a homogeneous population when examined by light microscopy or scanning electron microscopy. Morphologic and immunologic studies have been performed to characterize cultured endothelial cells (TABLE 1 ) . They have been unambiguously differentiated from possible contaminant cell types (medial smooth-muscle cells and fibroblasts) by the presence of: (1) A unique ultrastructural cytoplasmic marker, the Weibel-Palade body, (2) ABO blood group antigens appropriate to the endothelial cell donor's blood type, and (3) smooth-muscle actomyosin.
FIGURE 1. Photomicrograph of 5-day-old cultured endothelial cells. These cells are a homogeneous population of large, closely opposed, polygonal cells. (From Jaffe and colleagues I' by permission of the publishers of the J . Clin. Invest. ).
tmmunofluorescence Studies Studies were performed using immunofluorescence to demonstrate the presence of VIIIAGx in cultured endothelial cells. When cultured endothelial cells were sequentially incubated with a monospecific rabbit antibody to human factor VIII l6 and fluorescein-conjugated goat antirabbit gamma globulin, the cells fluoresced brilliantly (FIGURE 2). Similar studies performed on cultured fibroblasts and smooth-muscle cells were negative. Cultured endothelial cells did not fluoresce when normal rabbit serum was used instead of rabbit antifactor VIII. 1 3 9
Annals New York Academy of Sciences
64
TABLE1 DIFFERENTIATING CHARACTERISTICS OF ENDOTHELIAL CELLS *
WeibelWade bodies
cell type
Cell morphology
Growth pattern
Endothelial cell
Polygonal
Single layer N o pattern
Vascular smooth muscle cell Fibroblast
Spindle
Multilayered Parallel arrays
-
Spindle
Multilayered Parallel arrays Whorled
-
* From Jaffe and colleagues”
Immunofluorescent staining Presence with anti- of ABO smoothblood Myofilamuscle group ments actomyosin antigens
+ +
+
+
or variable
+ +
+
or variable
-
+
-
by permission of the publishers of the J. Clin. Invest.
FIGURE2. Immunofluorescence study of cultured human endothelial cells. Cells were treated with rabbit antifactor VIII and then with fluorescein-conjugated goat antirabbit gamma globulin. The cells are brightly stained. (From Jaffe and colleagues l’ by permission of the publishers of the J. Clin. Invest.).
65
JafTe: Synthesis of Factor VIII Antigen Demonstration and Partial Characterization of VIIIAus in Culture Medium and Cells
In these studies V I I I A G S was measured in preculture media, postculture media (after 72-hour incubations with cultured cells), and cell fractions by a sensitive radioimmunoassay.16 No VIIIAGs was present in the culture medium 2 ) ; VIIIAGS was detected in postculture before incubation with cells (TABLE media derived from endothelial cells; none was found in post culture media from either fibroblasts or smooth-muscle cells. The release of VIIIAG, into
TABLE2 AHF ANTIGEN CONTENT OF CULTURED CELLS Cell Type
AND
Material Tested
MEDIA * AHF Antigen units/ IOOml
Endothelial Cells
Smooth Muscle Cells
Fibroblasts
Preculture medium (3) f Postculture medium ( 3 ) Disrupted Cells 5 Supernatant ( 3 ) Pellet ( 3 ) Preculture medium ( 2 ) Postculture medium ( 3 ) Disrupted Cells 5 Supernatant ( 1 ) Pellet ( I ) Preculture medium (2) Postculture medium ( 3 )
< 0.2 1.4 2 1.4 %
16.2 2 1.4 13.6 & 1.9 < 0.3 < 0.3
< 1.0 < 1.0 < 0.3
< 0.3
* From Jatie and colleagues"
by permission of the publishers of the J. Clin. Invest. -i- Number of separate experiments.
t Mean 2 SEM
x 10' cells/ml) was frozen and thawed four times prior to centrifugation. The pellet was subsequently suspended in a volume of buffered saline equal to that of the supernatant fluid. 5 Cell suspension ( I S
the endothelial cell culture medium was linear over a 72-hour period. Cultured endothelial cells were separated into soluble and particulate material by freezethawing and centrifugation at 40,000 g. VIIIAGS was present in both soluble and particulate fractions from endothelial cells; no antigen was present in similar fractions prepared from cultured smooth-muscle cells. The physicochemical properties of endothelial VIIIAGs were like those of plasma VIIIAGS. When endothelial cell culture media was concentrated and chromatographed on Sepharose 6B, endothelial VIIIAG, eluted in the void volume, as did plasma V I I I A G S . Both endothelial cell VIII,G, and plasma VIIIAG, resisted adsorption by BaSO,, heating to 56" for 30 minutes, or incubation with bovine thrombin."
Annals New
66
York Academy of Sciences
Protein Synthesis by Cultured Endothelial Cells
These studies were performed to demonstrate the synthesis of protein by cultured endothelial cells. Cells incubated with 3H-leucine at 37" C incorporated radioactivity into TCA-precipitable protein. Incorporation was a linear function of both time and 3H-leucine dosage. Labeled TCA-precipitable material derived from cultured endothelial cells and their culture media was 95% solubilized by incubation with Pronase@, thus establishing that the 3H-leucine was incorporated into protein. Amino terminal addition of 3H-leucine to preformed protein was ruled out by amino-terminal amino acid analysis with dinitrofluorobenzene. Endothelial cell protein synthesis was inhibited by 93 % in the presence of cycloheximide (100 pgIml) . Synthesis of VIIIAas by Cultured Endothelial Cells
To demonstrate that cultured endothelial cells synthesize VIIIAGK rather than just release antigen previously accumulated in vivo, cultured endothelial cells were incubated with 3H-leucine. High-molecular-weight, VIIIAGs-rich fractions were prepared from the culture medium by ammonium sulfate precipitation and Sepharose 6B chromatography. Immunoprecipitable 3H-labeled protein was determined by a double antibody precipitin reaction using, first, either rabbit antifactor VIII or normal rabbit globulin and, second, goat antirabbit gamma g1obulin.l7 Seven percent of the radioactivity present in highmolecular-weight fractions prepared from endothelial cells was incorporated into immunoprecipitates formed by rabbit antifactor VIII (TABLE 3 ) . In contrast, only 1.5% of the 3H counts were coprecipitated by normal rabbit serum. Thus, approximately 5.5 % of 3H-leucine present in the high-molecular fraction was specifically incorporated into protein containing VIIIAGN. No specific precipitation of labeled protein was detected when similar fractions from smooth-muscle cell media were tested. The incorporation of 3H-leucine into VIIIAGN was suppressible by cycloheximide. Absence of V l l l A H 8Activity .
Endothelial cell postculture medium was assayed for VIIIAHFby a one-stage method 1s using AHF-deficient human plasma as substrate. Endothelial cells
TMMUNOPRECIPITATION
TABLE 3 'H LEUCINE-LABELED PROTEINS
OF
~
% cpm in Immunoprecipitates
Source of Culture Medium
Endothelial Cells Smooth-MuscleCells #'
Mean
&
SEM.
Rabbit Antifactor VIll 7.0 & 0.9 * 1.8
Control Rabbit Globulin 1.5 & 0.3 1.6
Jaffe: Synthesis of Factor VIII Antigen VIIIbH, OF
67
TABLE4 CULTURE MEDIUM
-
Material +
VIIIAHP units/100 ml
Medium alone Medium + EC t Medium + exogenous factor VIIl (0.7 u/ml) 5 Medium + EC exogenous factor VIII (0.7 u/ml) Medium + exogenous factor VIII (0.07 u/ml) Medium + EC + exogenous factor VIII (0.07 u/ml)
+
0.5 0.5 107
68
VIIIAOS units/100 ml 0.5
7.1 NT i l NT
4.5
NT
3.4
NT
* (From Jaffe et al?:
by permission of the publishers of the J. Clin. Invest.) t These components were incubated together for 72 hours and the supernatant culture medium assayed. t EC-cultured human endothelial cells. 5 Exogenous Factor VIII-Hemofil (Hyland Div.. Travenol Labs.). 11 NT-not tested.
cultured for 72 hours did not release VIIIAHF into their culture medium (TABLE 4). However, the endothelial cells did release VIIIAGs. VIII-iHF and VIIIAGs were both absent in preculture media. To demonstrate that newly synthesized VIII,El, would not be inactivated in the culture system, culture medium containing exogenous factor VIII (0.7 or 0.07 units/ml) was either incubated alone or with endothelial cells and then assayed for VIIIAHF. After 72 hours of incubation, VIIIAiHFwas present in the media at levels consistent with the amounts of exogenous factor VIII added (TABLE 4 ) . In a separate set of similar experiments (not shown) it was demonstrated that VIIIAc,s levels remained constant during the incubation period. Thus, neither VIIIAGs nor VIII.lHF were significantly degraded during the incubation period. Attempts to activate VIIIAFIpin postculture medium with thrombin were unsuccessful.
DISCUSSION These studies demonstrate that cultured human endothelial cells synthesize and release VIIIA,:s. These observations are compatible with the recent demonstration of VIIIA(;s in endothelial cells throughout the body.11-13 Control immunofluorescence and radioimmunoassay studies established that cultured smooth-muscle cells and fibroblasts do not have this capacity. In contrast to the secretion of VIIIAGs, no VIIIAH, has been detected in the postculture medium. There are several possible explanations for this finding: ( 1 ) It is possible that endothelial cells in various anatomic regions are functionally heterogeneous. ( 2 ) The artificial conditions of the culture might affect VIIIA(;s synthesis in a way that would prevent the development of VIIIAIIFq(3) Endothelial cells may synthesize an inactive “precursor” mole-
Annals New
68
York Academy of Sciences
cule. If this is the case, the conversion necessary for VIIIAHp activity might occur at a remote site by a different cell type and thus not be possible in tissue culture. (4) Endothelial cells may normally synthesize only one component of a macromolecular complex that has VIIIAHF. Several laboratories have recently identified factor VIII subunits when this protein is dissociated by reduction and alkylation l. 18* ?O or by high salt c~ncentration.?~-*~ When factor VIII is dissociated by 0.24M CaCl, and chromatographed on agarose gel columns, all the VIIIAQS and none of the VIIIAHF is found in the high-molecular-weight fractions. Conversely, all the VIIIAH, but none of the VIIIAGN is found in the low-molecular-weight fractions.?-’ However, human and rabbit antibodies that inactivate VIIIAHF are neutralized by antigenic determinants on the highmolecular-weight The presence of common antigens on the two subunits suggests a molecular relationship that remains to be defined.
SUMMARY Endothelial cells have been isolated from freshly obtained human umbilical cords by collagenase digestion of the interior of the umbilical vein and grown in tissue culture. These cells have been identified by morphologic and immunologic criteria. It has been demonstrated that these cultured human endothelial cells synthesize and release factor VIII antigen but not factor VIII clot-promoting activity.
REFERENCES
E. W. DAVIE.1973. Isolation and characterization of human factor VIII (antihemophilic factor). J. Biol. Chem.
1. LEGAZ,M. E., G. SCHMER, R. B. COUNTS &
2 4 8 3946. 2. WEISS,H. J., L. W.HOYER, F. R. RICKLES, A. VARMA & J. ROGERS.1973. Quantitative assay of a plasma factor deficient in von Willebrand’s disease that is necessary for platelet aggregation. J. Clin. Invest. 52: 2708. 3. NORMAN, J. C., J. P. LAMBILLIOTE, Y.KOZIMA& H. S. SISE. 1967. Antihemo4.
5. 6.
7. 8.
9. 10.
philic factor release by perfused liver and spleen: relationship to hemophilia. Science 1 5 8 1060. DODDS, W. J. 1969. Storage, release and synthesis of coagulation factors in isolated perfused organs. Amer. J. Physiol. 217: 879. WEBSTER, W. P., C. F. ZUKOSKI, P. HUTCHINS, R. L. REDDICK, S. R. MANDEL& G. D. PENICK.1971. Plasma factor VIII synthesis and control as revealed by canine organ transplantation. Amer. J. Physiol. 2 2 0 1147. POOL,J. G. 1966. Antihemophilic globulin (AHG, factor VIII) activity in spleen. Fed. Proc. 25: 317. (Abstr.). WEBSTER, W. P. & G. D. PENICK.1967. Release of factor VIII from isolated perfused organs. Fed. Proc. 26: 487. (Abstr.). BARROW,E. M. & J. B. GRAHAM. 1968. Kidney antihemophilic factor. Partial purification and some properties. Biochemistry 7: 3917. POOL,J. G. & T.H. SPAET.1954. Ethionine-induced depression of plasma antihemophilic globulin in the rat. Proc. Soc. Exp. Biol. Med. 87: 54. BOUHASIN,J. D., P. MONTELEONE & C. ALTAY.1971. Role of the lymphocyte in antihemophilic globulin production: a rise in antihemophilic globulin levels in a hemophilic subject with acute lymphoblastic leukemia. J. Lab. Clin. Med. 7 8 122.
Jaffe: Synthesis of Factor VIII Antigen 11.
12. 13. 14. 15. 16.
69
SANTOS,R. P. & L. W. HOYER.1972. Antihemophilic factor in tissues: localization by immunofluorescence. Fed. Proc. 31: 262. (Abstr.). BLOOM,A. L., J. C. GIDDINGS & C. J. WILKS. 1973. Factor VIII on the vascular intima: possible importance in hemostasis and thrombosis. Nat. New Biol. 241: 217. HOYER,L. W., R. P. DE LOS SANTOS & J. R. HOYER. 1973. Antihemophilic factor antigen. Localization in endothelial cells by immunofluorescent microscopy. J. Clin. Invest. 52: 2737. JAFFE,E. A., R. L. NACHMAN & C. G. BECKER.1972. Culture of human endothelial cells derived from umbilical cord veins. J. Clin. Invest. 51: 46a (Abstr). JAFFE,E. A., R. L. NACHMAN, C. G. BECKER& C. R. MINICK.1973. Culture of human endothelial cells derived from umbilical cord veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52: 2745. HOYER,L. W. 1972. Immunologic studies of antihemophilic factor (AHF, factor VIII). IV. Radioimmunoassay of AHF antigen. J. Lab. Clin. Med. 8 0 DE LOS
822. 1973. Synthesis of antihemophi17. JAFFE,E. A., L. W. HOYER& R. L. NACHMAN.
lic factor antigen by cultured human endothelial cells. J. Clin. Invest. 52: 2757. R. T. & 0. D. RATNOFF.1962. Studies on the nature of the cir18. BRECKENRIDGE, 19. 20. 21. 22. 23. 24.
culating anticoagulant directed against antihemophilic factor: with notes on an assay for antihemophilic factor. Blood 2 0 137. ANDERSON, J .C. & P. A. MCKEE. 1972. The effects of proteolytic enzymes on the coagulant properties and molecular structure of human factor VIII. Circulation 46(Suppl. 11): 52 (Abstr.). S. L., N. R. SHULMAN & H. R. GRALNICK. 1972. Studies on the puriMARCHESI, fication and characterization of human factor VIII. J. Clin. Invest. 51: 2151. WEISS,H. J. & S. KWHWA. 1970. Molecular forms of antihaemophilic globulin in plasma, cryoprecipitate and after thrombin activation. Brit. J. Haematol. 18: 89. OWEN,W. G. & R. H. WAGNER.1972. Antihemophilic factor: separation of an active fragment following dissociation by salts or detergents. Thromb. Diath. Haemorrh. 27: 502. WEISS,H. J., L. L. PHILUPS& W. ROSNER. 1972. Separation of subunits of antihemophilic factor (AHF) by agarose gel chromatography. Thromb. Diath. Hemorrh. 27: 212. RICK,M. E. & L. W. HOYER. 1973. Immunologic studies of antihemophilic factor (AHF, Factor VIII). V. Immunologic properties of AHF subunits produced by salt dissociation. Blood 42: 737.
Factor VIII (antihemophilic factor, AHF) is a plasma glycoprotein that in the purified state has a molecular weight of 1.12 x 1OS.l The partially purified protein has three distinct properties: (a) factor VIII antigen (VIIIAGX), an antigen identified in precipitin assays by heterologous antibodies and decreased in plasma of patients with von Willebrand’s disease but normal in hemophilic plasma; (b) factor VIII clot promoting function (VIIIAHF), an activity that corrects the plasma coagulation abnormalities of patients with classic hemophilia (hemophilia A ) ; and (c) von Willebrand factor (VIIIAHF), an activity deficient in von Willebrand‘s disease, which can be identified by an abnormality of the bleeding time or by in vitro assays of platelet function (retention of platelets in glass bead columns o r ristocetin-induced aggregation of washed normal human platelets). The term “factor VIII” identifies a protein(s) present in normal human plasma which is (are) responsible for these three activities. This operational nomenclature has recently been suggested by Weiss and coworkers.2 The site of productions of factor VIII is unknown. The l i ~ e r , ~ - ~ spleen,3* 6 , kidney,x reticuloendothelial cell: and lymphocytes lo have all been suggested as sites of synthesis. Recent immunofluorescence studies performed on a variety of human tissue sections, using a monospecific rabbit antihuman factor VIII, have demonstrated that endothelial cells contain VIIIAGs.11-13The development in our laboratory of methods for culturing and identifying human endothelial cells l 4 3 1s has made it possible to examine the role of the endothelial cell in factor VIII synthesis. Our results show that cultured human endothelial cells synthesize and release VIIIAGs into their culture medium. 4v
RESULTS Culture of Human Endothelial Cells
Endothelial cells were obtained from human umbilical cord veins by digestion of the vein’s interior with 0.2% collagenase for 15 minutes at 37” C. The collagenase solution was drained from the vein, which was then rinsed with cord buffer (0.14 M NaCI, 0.004 M KCl, 0.001 M phosphate buffer, pH 7.4, 0.01 1 M glucose). The cells present in these solutions were sedimented by centrifugation at 250 g, washed once with Medium 199 containing 20% :ii Supported by grants from the National Institutes of Health (HL-14810)and The Arnold R. Krakower Hematology Foundation.
62
Jaffe: Synthesis of Factor VIII Antigen
63
fetal calf serum, and cultured in this medium at 37" under 5 % C02.15 Smoothmuscle cells and fibroblasts were also cultured as previously described.15 After one week in culture, endothelial cells grew in confluent monolayers without a definable whorling pattern. The cells were large (30 x 50 p ) , polygonal, and closely apposed with an oval, centrally located nucleus and indistinct cell borders (FIGURE1). The cells formed a homogeneous population when examined by light microscopy or scanning electron microscopy. Morphologic and immunologic studies have been performed to characterize cultured endothelial cells (TABLE 1 ) . They have been unambiguously differentiated from possible contaminant cell types (medial smooth-muscle cells and fibroblasts) by the presence of: (1) A unique ultrastructural cytoplasmic marker, the Weibel-Palade body, (2) ABO blood group antigens appropriate to the endothelial cell donor's blood type, and (3) smooth-muscle actomyosin.
FIGURE 1. Photomicrograph of 5-day-old cultured endothelial cells. These cells are a homogeneous population of large, closely opposed, polygonal cells. (From Jaffe and colleagues I' by permission of the publishers of the J . Clin. Invest. ).
tmmunofluorescence Studies Studies were performed using immunofluorescence to demonstrate the presence of VIIIAGx in cultured endothelial cells. When cultured endothelial cells were sequentially incubated with a monospecific rabbit antibody to human factor VIII l6 and fluorescein-conjugated goat antirabbit gamma globulin, the cells fluoresced brilliantly (FIGURE 2). Similar studies performed on cultured fibroblasts and smooth-muscle cells were negative. Cultured endothelial cells did not fluoresce when normal rabbit serum was used instead of rabbit antifactor VIII. 1 3 9
Annals New York Academy of Sciences
64
TABLE1 DIFFERENTIATING CHARACTERISTICS OF ENDOTHELIAL CELLS *
WeibelWade bodies
cell type
Cell morphology
Growth pattern
Endothelial cell
Polygonal
Single layer N o pattern
Vascular smooth muscle cell Fibroblast
Spindle
Multilayered Parallel arrays
-
Spindle
Multilayered Parallel arrays Whorled
-
* From Jaffe and colleagues”
Immunofluorescent staining Presence with anti- of ABO smoothblood Myofilamuscle group ments actomyosin antigens
+ +
+
+
or variable
+ +
+
or variable
-
+
-
by permission of the publishers of the J. Clin. Invest.
FIGURE2. Immunofluorescence study of cultured human endothelial cells. Cells were treated with rabbit antifactor VIII and then with fluorescein-conjugated goat antirabbit gamma globulin. The cells are brightly stained. (From Jaffe and colleagues l’ by permission of the publishers of the J. Clin. Invest.).
65
JafTe: Synthesis of Factor VIII Antigen Demonstration and Partial Characterization of VIIIAus in Culture Medium and Cells
In these studies V I I I A G S was measured in preculture media, postculture media (after 72-hour incubations with cultured cells), and cell fractions by a sensitive radioimmunoassay.16 No VIIIAGs was present in the culture medium 2 ) ; VIIIAGS was detected in postculture before incubation with cells (TABLE media derived from endothelial cells; none was found in post culture media from either fibroblasts or smooth-muscle cells. The release of VIIIAG, into
TABLE2 AHF ANTIGEN CONTENT OF CULTURED CELLS Cell Type
AND
Material Tested
MEDIA * AHF Antigen units/ IOOml
Endothelial Cells
Smooth Muscle Cells
Fibroblasts
Preculture medium (3) f Postculture medium ( 3 ) Disrupted Cells 5 Supernatant ( 3 ) Pellet ( 3 ) Preculture medium ( 2 ) Postculture medium ( 3 ) Disrupted Cells 5 Supernatant ( 1 ) Pellet ( I ) Preculture medium (2) Postculture medium ( 3 )
< 0.2 1.4 2 1.4 %
16.2 2 1.4 13.6 & 1.9 < 0.3 < 0.3
< 1.0 < 1.0 < 0.3
< 0.3
* From Jatie and colleagues"
by permission of the publishers of the J. Clin. Invest. -i- Number of separate experiments.
t Mean 2 SEM
x 10' cells/ml) was frozen and thawed four times prior to centrifugation. The pellet was subsequently suspended in a volume of buffered saline equal to that of the supernatant fluid. 5 Cell suspension ( I S
the endothelial cell culture medium was linear over a 72-hour period. Cultured endothelial cells were separated into soluble and particulate material by freezethawing and centrifugation at 40,000 g. VIIIAGS was present in both soluble and particulate fractions from endothelial cells; no antigen was present in similar fractions prepared from cultured smooth-muscle cells. The physicochemical properties of endothelial VIIIAGs were like those of plasma VIIIAGS. When endothelial cell culture media was concentrated and chromatographed on Sepharose 6B, endothelial VIIIAG, eluted in the void volume, as did plasma V I I I A G S . Both endothelial cell VIII,G, and plasma VIIIAG, resisted adsorption by BaSO,, heating to 56" for 30 minutes, or incubation with bovine thrombin."
Annals New
66
York Academy of Sciences
Protein Synthesis by Cultured Endothelial Cells
These studies were performed to demonstrate the synthesis of protein by cultured endothelial cells. Cells incubated with 3H-leucine at 37" C incorporated radioactivity into TCA-precipitable protein. Incorporation was a linear function of both time and 3H-leucine dosage. Labeled TCA-precipitable material derived from cultured endothelial cells and their culture media was 95% solubilized by incubation with Pronase@, thus establishing that the 3H-leucine was incorporated into protein. Amino terminal addition of 3H-leucine to preformed protein was ruled out by amino-terminal amino acid analysis with dinitrofluorobenzene. Endothelial cell protein synthesis was inhibited by 93 % in the presence of cycloheximide (100 pgIml) . Synthesis of VIIIAas by Cultured Endothelial Cells
To demonstrate that cultured endothelial cells synthesize VIIIAGK rather than just release antigen previously accumulated in vivo, cultured endothelial cells were incubated with 3H-leucine. High-molecular-weight, VIIIAGs-rich fractions were prepared from the culture medium by ammonium sulfate precipitation and Sepharose 6B chromatography. Immunoprecipitable 3H-labeled protein was determined by a double antibody precipitin reaction using, first, either rabbit antifactor VIII or normal rabbit globulin and, second, goat antirabbit gamma g1obulin.l7 Seven percent of the radioactivity present in highmolecular-weight fractions prepared from endothelial cells was incorporated into immunoprecipitates formed by rabbit antifactor VIII (TABLE 3 ) . In contrast, only 1.5% of the 3H counts were coprecipitated by normal rabbit serum. Thus, approximately 5.5 % of 3H-leucine present in the high-molecular fraction was specifically incorporated into protein containing VIIIAGN. No specific precipitation of labeled protein was detected when similar fractions from smooth-muscle cell media were tested. The incorporation of 3H-leucine into VIIIAGN was suppressible by cycloheximide. Absence of V l l l A H 8Activity .
Endothelial cell postculture medium was assayed for VIIIAHFby a one-stage method 1s using AHF-deficient human plasma as substrate. Endothelial cells
TMMUNOPRECIPITATION
TABLE 3 'H LEUCINE-LABELED PROTEINS
OF
~
% cpm in Immunoprecipitates
Source of Culture Medium
Endothelial Cells Smooth-MuscleCells #'
Mean
&
SEM.
Rabbit Antifactor VIll 7.0 & 0.9 * 1.8
Control Rabbit Globulin 1.5 & 0.3 1.6
Jaffe: Synthesis of Factor VIII Antigen VIIIbH, OF
67
TABLE4 CULTURE MEDIUM
-
Material +
VIIIAHP units/100 ml
Medium alone Medium + EC t Medium + exogenous factor VIIl (0.7 u/ml) 5 Medium + EC exogenous factor VIII (0.7 u/ml) Medium + exogenous factor VIII (0.07 u/ml) Medium + EC + exogenous factor VIII (0.07 u/ml)
+
0.5 0.5 107
68
VIIIAOS units/100 ml 0.5
7.1 NT i l NT
4.5
NT
3.4
NT
* (From Jaffe et al?:
by permission of the publishers of the J. Clin. Invest.) t These components were incubated together for 72 hours and the supernatant culture medium assayed. t EC-cultured human endothelial cells. 5 Exogenous Factor VIII-Hemofil (Hyland Div.. Travenol Labs.). 11 NT-not tested.
cultured for 72 hours did not release VIIIAHF into their culture medium (TABLE 4). However, the endothelial cells did release VIIIAGs. VIII-iHF and VIIIAGs were both absent in preculture media. To demonstrate that newly synthesized VIII,El, would not be inactivated in the culture system, culture medium containing exogenous factor VIII (0.7 or 0.07 units/ml) was either incubated alone or with endothelial cells and then assayed for VIIIAHF. After 72 hours of incubation, VIIIAiHFwas present in the media at levels consistent with the amounts of exogenous factor VIII added (TABLE 4 ) . In a separate set of similar experiments (not shown) it was demonstrated that VIIIAc,s levels remained constant during the incubation period. Thus, neither VIIIAGs nor VIII.lHF were significantly degraded during the incubation period. Attempts to activate VIIIAFIpin postculture medium with thrombin were unsuccessful.
DISCUSSION These studies demonstrate that cultured human endothelial cells synthesize and release VIIIA,:s. These observations are compatible with the recent demonstration of VIIIA(;s in endothelial cells throughout the body.11-13 Control immunofluorescence and radioimmunoassay studies established that cultured smooth-muscle cells and fibroblasts do not have this capacity. In contrast to the secretion of VIIIAGs, no VIIIAH, has been detected in the postculture medium. There are several possible explanations for this finding: ( 1 ) It is possible that endothelial cells in various anatomic regions are functionally heterogeneous. ( 2 ) The artificial conditions of the culture might affect VIIIA(;s synthesis in a way that would prevent the development of VIIIAIIFq(3) Endothelial cells may synthesize an inactive “precursor” mole-
Annals New
68
York Academy of Sciences
cule. If this is the case, the conversion necessary for VIIIAHp activity might occur at a remote site by a different cell type and thus not be possible in tissue culture. (4) Endothelial cells may normally synthesize only one component of a macromolecular complex that has VIIIAHF. Several laboratories have recently identified factor VIII subunits when this protein is dissociated by reduction and alkylation l. 18* ?O or by high salt c~ncentration.?~-*~ When factor VIII is dissociated by 0.24M CaCl, and chromatographed on agarose gel columns, all the VIIIAQS and none of the VIIIAHF is found in the high-molecular-weight fractions. Conversely, all the VIIIAH, but none of the VIIIAGN is found in the low-molecular-weight fractions.?-’ However, human and rabbit antibodies that inactivate VIIIAHF are neutralized by antigenic determinants on the highmolecular-weight The presence of common antigens on the two subunits suggests a molecular relationship that remains to be defined.
SUMMARY Endothelial cells have been isolated from freshly obtained human umbilical cords by collagenase digestion of the interior of the umbilical vein and grown in tissue culture. These cells have been identified by morphologic and immunologic criteria. It has been demonstrated that these cultured human endothelial cells synthesize and release factor VIII antigen but not factor VIII clot-promoting activity.
REFERENCES
E. W. DAVIE.1973. Isolation and characterization of human factor VIII (antihemophilic factor). J. Biol. Chem.
1. LEGAZ,M. E., G. SCHMER, R. B. COUNTS &
2 4 8 3946. 2. WEISS,H. J., L. W.HOYER, F. R. RICKLES, A. VARMA & J. ROGERS.1973. Quantitative assay of a plasma factor deficient in von Willebrand’s disease that is necessary for platelet aggregation. J. Clin. Invest. 52: 2708. 3. NORMAN, J. C., J. P. LAMBILLIOTE, Y.KOZIMA& H. S. SISE. 1967. Antihemo4.
5. 6.
7. 8.
9. 10.
philic factor release by perfused liver and spleen: relationship to hemophilia. Science 1 5 8 1060. DODDS, W. J. 1969. Storage, release and synthesis of coagulation factors in isolated perfused organs. Amer. J. Physiol. 217: 879. WEBSTER, W. P., C. F. ZUKOSKI, P. HUTCHINS, R. L. REDDICK, S. R. MANDEL& G. D. PENICK.1971. Plasma factor VIII synthesis and control as revealed by canine organ transplantation. Amer. J. Physiol. 2 2 0 1147. POOL,J. G. 1966. Antihemophilic globulin (AHG, factor VIII) activity in spleen. Fed. Proc. 25: 317. (Abstr.). WEBSTER, W. P. & G. D. PENICK.1967. Release of factor VIII from isolated perfused organs. Fed. Proc. 26: 487. (Abstr.). BARROW,E. M. & J. B. GRAHAM. 1968. Kidney antihemophilic factor. Partial purification and some properties. Biochemistry 7: 3917. POOL,J. G. & T.H. SPAET.1954. Ethionine-induced depression of plasma antihemophilic globulin in the rat. Proc. Soc. Exp. Biol. Med. 87: 54. BOUHASIN,J. D., P. MONTELEONE & C. ALTAY.1971. Role of the lymphocyte in antihemophilic globulin production: a rise in antihemophilic globulin levels in a hemophilic subject with acute lymphoblastic leukemia. J. Lab. Clin. Med. 7 8 122.
Jaffe: Synthesis of Factor VIII Antigen 11.
12. 13. 14. 15. 16.
69
SANTOS,R. P. & L. W. HOYER.1972. Antihemophilic factor in tissues: localization by immunofluorescence. Fed. Proc. 31: 262. (Abstr.). BLOOM,A. L., J. C. GIDDINGS & C. J. WILKS. 1973. Factor VIII on the vascular intima: possible importance in hemostasis and thrombosis. Nat. New Biol. 241: 217. HOYER,L. W., R. P. DE LOS SANTOS & J. R. HOYER. 1973. Antihemophilic factor antigen. Localization in endothelial cells by immunofluorescent microscopy. J. Clin. Invest. 52: 2737. JAFFE,E. A., R. L. NACHMAN & C. G. BECKER.1972. Culture of human endothelial cells derived from umbilical cord veins. J. Clin. Invest. 51: 46a (Abstr). JAFFE,E. A., R. L. NACHMAN, C. G. BECKER& C. R. MINICK.1973. Culture of human endothelial cells derived from umbilical cord veins. Identification by morphologic and immunologic criteria. J. Clin. Invest. 52: 2745. HOYER,L. W. 1972. Immunologic studies of antihemophilic factor (AHF, factor VIII). IV. Radioimmunoassay of AHF antigen. J. Lab. Clin. Med. 8 0 DE LOS
822. 1973. Synthesis of antihemophi17. JAFFE,E. A., L. W. HOYER& R. L. NACHMAN.
lic factor antigen by cultured human endothelial cells. J. Clin. Invest. 52: 2757. R. T. & 0. D. RATNOFF.1962. Studies on the nature of the cir18. BRECKENRIDGE, 19. 20. 21. 22. 23. 24.
culating anticoagulant directed against antihemophilic factor: with notes on an assay for antihemophilic factor. Blood 2 0 137. ANDERSON, J .C. & P. A. MCKEE. 1972. The effects of proteolytic enzymes on the coagulant properties and molecular structure of human factor VIII. Circulation 46(Suppl. 11): 52 (Abstr.). S. L., N. R. SHULMAN & H. R. GRALNICK. 1972. Studies on the puriMARCHESI, fication and characterization of human factor VIII. J. Clin. Invest. 51: 2151. WEISS,H. J. & S. KWHWA. 1970. Molecular forms of antihaemophilic globulin in plasma, cryoprecipitate and after thrombin activation. Brit. J. Haematol. 18: 89. OWEN,W. G. & R. H. WAGNER.1972. Antihemophilic factor: separation of an active fragment following dissociation by salts or detergents. Thromb. Diath. Haemorrh. 27: 502. WEISS,H. J., L. L. PHILUPS& W. ROSNER. 1972. Separation of subunits of antihemophilic factor (AHF) by agarose gel chromatography. Thromb. Diath. Hemorrh. 27: 212. RICK,M. E. & L. W. HOYER. 1973. Immunologic studies of antihemophilic factor (AHF, Factor VIII). V. Immunologic properties of AHF subunits produced by salt dissociation. Blood 42: 737.
