Alvin H. SchmaierSB, Susan C. Murrayll, Ghanshyam D. HedaS, Anthony FarberS, Alice Kuon,. Keith McCraen, and Douglas B. Cinesll. From the Departments of ...
Vol. 264, No. 30, Issue of October 25, pp. 18173-18179,1989 Printed in U.S A.
THEJOURNALOF BIOLOGICAL CHEMISTRY Q 1989 by The American Soclety for Biochemistry and Molecular Biology, Inc.
Synthesis andExpression of C 1 Inhibitor by Human Umbilical Vein Endothelial Cells* (Received for publication, January 31, 1989)
Alvin H. SchmaierSB,Susan C. Murrayll, Ghanshyam D. HedaS, Anthony FarberS,Alice Kuon, Keith McCraen, and Douglas B. Cinesll From the Departments of Medicine, Hematology/Oncology Sections, $Temple Uniuersity and the TUniuersity of Pennsylvania, Philadelphia, Pennsyluania 19140
The biologic activity of C1 esterase, activated forms The vessel wall is frequently exposed to proteolytic enzymes vascular tissue at of factor XI1 and kallikrein at sites of vascular inflam- that are generated in plasma and within mation may be regulated by C 1 inhibitor (C1 INH) sites of inflammation. The mechanisms by which vascular elaborated by endothelial cells. Therefore, we investi- cells themselves are protected fromproteolysis are incomgated whether human umbilical vein endothelial cells pletely defined. This protection may depend, in part, upon (HUVEC) in culture produce C1 INH. Passaged HU- the production and/or expression by endothelial cells and VEC contain 1.6 f 0.8 pg of C1 INH/108 cells (mean2 other cells within the vasculature of inhibitors of proteolytic S.D.; n = 7) which was immunochemically similar to enzymes. plasma C1 INH measured by a competitive enzymeEndothelial cell injury is frequently associated with actilinked immunosorbent assay. Methylamine-treated ly- vation of complement and prekallikrein (1,2). Human umbilsates of HUVEC contained a functional inhibitor of ical vein endothelial cells (HUVEC)’ in culture express surpurified kallikrein (2.7 2 0.8 pg activity/lO*cells, mean f S.D.; n = 4). The HUVEC-derived kallikrein face receptors for C l q whichmayserve to localize the C1 inhibitory activity was mostly C 1 INH because it was esterase complex (3). HUVEC also both synthesize and exreversed by chemically treating the lysate with chlo- press surface-binding sitesfor high molecular weight kininoroform and was neutralizedby a n t i 4 1 INH antibody. gen (4,5 ) which can be cleaved by kallikrein and activated forms of factor XI1 to liberate the vasoactive peptide bradyA lysate of HUVEC derived from an umbilical cord fromapatientwithType I hereditary angioedema kinin (6-9),a potentstimulus of several endothelial cellcontained less than 30%of the normallevels of C1 INH derived processes (10-12). Inordertomodulatecomplementactivationandkinin antigen and activity. Immunohistochemical staining of HUVEC demonstrated a diffuse pattern of staining for formation, endothelial cells may also express inhibitors that of C1 esterase, activated forms C 1 INH. HUVEC C 1 INH was also expressed on the regulate the enzymatic activity endothelial cell surfaceas detected by binding of anti- of factor XI1 and kallikrein. C1 inhibitor (Cl INH) is the C1 INH antibody to intact monolayers and was elabo- only known plasma protease inhibitorof C1 esterase and is a rated progressively into the overlyingmedia over the major plasma protease inhibitor of the enzymes of contact first 24 hinculture. HUVEC incubatedwith [36S] system, activated forms of factor XI1 and kallikrein (13-16). methionine secreted a metabolically labeled protein Therefore, we investigatedwhetherHUVEC produce and having a molecular mass of 92 kDa immunoisolated express a functionally active form of C1 INH. using polyclonal or monoclonal antibodies to human1C INH. A mRNA transcript encoding for C1 INH was EXPERIMENTALPROCEDURES detected by slot blot hybridization. Incubation of HUMaterials-Goat antisera monospecific for humanC1 INH was VEC with y-interferon stimulated the expression of the 2.1 kilobasemRNA for C1INH and increased the level purchased from Atlantic Antibodies (Scarborough, ME). Incubation of this antiserawith plasmaproduced a single precipitin arc ondouble of C1 INH produced by these cells. Production and immunodiffusion which showed complete identity with purified C1 expression of C 1 INH by endothelial cells may help as previously reported (17). Mouse IgG in ascites (NS-1) was modulate the complement system and the contactsys- INH, obtained from Organon-Teknika (Malvern, PA). A rabbitanti-human tem of plasma proteolysis on the vascular surface in von Willebrandfactorantibody was purchased from Dako Corp. vivo. (Santa Barbara, CA). Na-lZ5I (50 mCi/mmol) and [35S]methionine ~~
~~~
~
~
~
~
* This work wassupported in part by HL35553 (to A. H. S.),GrantIn-Aid 861169 from the American Heart Association (to A. H. s.),a Grant-In-Aid from the Southeastern Pennsylvania Affiliate of the American Heart Association (to A. H. S.),Grant 6437 from the March of Dimes Birth Defect Foundation (to A. H. S.), HL34044 (to D. B. C.), and Grant 2381 from The Tobacco Research Council (to D. B. C.) The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. §Recipient of Research Career DevelopmentAward HL01615 from the National Institutes of Health. To whom reprint requests should be addressed Hematology/Oncology Section,TempleUniversity School of Medicine, 3400 N. Broad St., Philadelphia, PA 19140.
(1000 Ci/mmol) were purchased from Amersham Corp. [32P]dATP (3000 Ci/mmol) was obtained from ICN Pharmaceuticals(Irvine, CA). 12sII-Labeledgoat anti-mouse IgG was purchased from Du PontNew England Nuclear. Iodogen (chloroamide, 1,3,4,6-tetrachloro-3a, 6a-diphenylglycoluril) was obtained from Pierce Chemical Co. Nitrocellulose (0.45-pm poresize), was purchased from Schleicher & Schuell. All reagent and molecular weight standards for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) were purchased from Bio-Rad. Pansorbin was purchased from Behring Diagnostics. H-D-Pro-Phe-Arg-pNA (S2303) chromogenic substrate was purchased from Helena Laboratories (Beaumont, TX). ChloroThe abbreviations used are: HUVEC, human umbilical vein endothelial cells; C1 INH, C1 inhibitor; CELISA, competitive enzymelinked immunosorbent assay; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; IFN-y, y-interferon; SSC, standard sodium citrate; kb, kilobase.
18173
HUVEC C1 INH RESULTS
Identification of HUVEC C l INH Antigen-We first determined whether passaged HUVEC contained C1INH antigen. Lysates prepared from multiply passaged HUVEC contained 1.6 & 0.8 pg C1 INH/lOS cells (mean f S.D.) (range 0.63 to 2.7 pg/108 cells) determined by CELISA (Table I). C1 INH identified in the endothelial cell lysate and in normal plasma were compared using the CELISA. The shape of the competition inhibition curves generated by equal amounts of C1 INH antigen from both sources were superimposible (Fig. l), indicating that C1 INH antigen from each source was recognized equally by the polyclonal antibody to human C1 INH. Moreover, a lysate preparedfrom multiply passaged HUVEC derived from a woman with Type I hereditary angioedema TABLE I C1 1NH content of HUVEC lysates pg
Sample
C1 INH/108 cells"
HUVEC C1 INH antigen Kallikrein inhibition by HUVEC lysate ( N = 4)b Inhibition of kallikrein by HUVEC lysate incubated with chloroform ( N = 3)' Inhibition of kallikrein by HUVEC lysate incubated with anti-C1 INH antibody ( N = 2)d
1.6 f 0.8 2.7 0.8
*
0.03 + 0.02 0.14
The numbersrepresent the mean f S.D. amount of C1 INH antigen or kallikrein neutralizing activity measured in the HUVEC lysates. All samples for functional studies were incubated with methylamine (see "Experimental Procedures") prior to determination of kallikrein inhibitory activity. b T h e number in parenthesis indicates the number of separate experiments performed. Chloroform treatment consisted of mixing the methylaminetreated HUVEC lysate with an equal volume of ice-cold chloroform as previously reported (21, 35). The lysate was then vortexed for 1 min at room temperature and centrifuged a t 12,000 X g for 4 min a t 4 "C. The residual kallikrein neutralizing activity in the supernatant was then determined. A lysate of HUVEC was incubated with antibody to C1 INH for 16 b at 4 "C. The 1ysat.e was treated withmethylamine and the residual kallikrein neutralizing activity was measured.
1.0.0
0
0
-
0 ENDOTHELIAL CELLS
A NHP
.6-
V 0
0
a .4.2
I
I
I
.5
I
I
I
I
2.5 5 IO ng C ~ I N H
I
I
25
50
100
FIG. 1. Comparison of HUVEC and plasma C1 INH antigen. Various concentrations of HUVEC-derived C1 INH (0)and plasma C l INH (A) were compared using the competitive ELISA. Rcalc OD represents the relative degree (between 1 and 0) of residual antibody after being incubated with the C1 INH. Thefigure is a representative experiment of two performed.
18175
grown in similar medium had a C1 INH antigen level of 0.19 pg/108 cells. In contrast,lysates from pooled normal endothelial cells grown in the same medium for the same number of passages contained 2.2 pg/lOs cells. Determination of HUVEC Cl INH Activity-Additional investigations were performed to determine if the C1 INH in HUVEC lysates was functionally active. HUVEC lysates contained a total amountof kallikrein neutralizing activity equivalent to 2.7 f 0.8 pg of C1 INH/1OS cells (mean S.D.; n = 4) (Table I). This value was similar to theamount of C1 INH antigen detectedby CELISA. Further studies were performed to determine if the kallikrein neutralizing activity of the HUVEC lysates was due to the presence of C1 INH (Table I). When the same HUVEC lysates were pretreated with chloroform, a reversible inhibitor of C1 INH (35), the capacity of the HUVEC lysate to inhibit kallikrein's amidolytic activity was decreased 70-fold to a mean level of activity equivalent to 0.03 pg of C1 INH/1OS cells. When lysates from HUVEC were preincubated with an antibody to C1 INH, thekallikrein inhibitory activity was reduced 19-fold to the equivalent of 0.14pgof C1 INH/1OScells. Finally, lysates of HUVEC derived from an umbilical cord from a patient with Type I hereditary angioedema contained little,if any, kallikrein neutralizing activity. Together these combined studies indicate that the kallikrein inhibitory activity in endothelial cell lysates was due primarily to the presence of C1 INH. Distribution of HUVEC C l INH Antigen-C1 INH antigen was readily detected in HUVEC by immunohistochemistry. C1 INH appeared to be present in a diffuse pattern in permeabilized cells, similar to Von Willebrand factor (Fig. 2). When a control murine IgG (Fig. 2) or a control rabbit IgG (data notshown) were used, no staining was seen. In addition, the presence of C1 INH antigen onthe cell surface of HUVEC was evaluated using a radioimmunoassay. Greater than 4-fold more monoclonal anti-Cl INHantibody bound to monolayers of HUVEC than did the same concentrationof control murine IgG, indicating that atleast part of the totalHUVEC C1 INH was expressed on the HUVEC surface (data not shown). Additional studies were performed to determine if HUVEC C1 INH was secreted into the conditioned media (Fig. 3). HUVEC (passages 2 or 5) were washed and serum-free media was added to thecell cultures for 1, 2,6, and24 h. The media was then removed and the concentration of C1 INH antigen was determined by CELISA (Fig. 3A). The concentration of C1 INH in themedia increased progressively from a concentration of 132 ng/ml at 1 h to a concentration of 229 ng/ml at 24 h. The effect of cell passage on the amount of C1 INH secreted intothe media was also studied (Fig. 3B).The concentration of C1 INH in conditioned media from cells passaged five or six times were significantly higher ( p < 0.03) than thatpresent in media from HUVEC passaged once. Synthesis of C1 INH by HUVEC-The C1 INH antigen present in HUVEC lysates on their surface and secreted into the conditioned media was not derived from the fetal calf serum in which the cells weregrown. Undiluted fetal calf serum contained a concentration of C1 INH of
