[CANCER RESEARCH 48, 3507-3514, June 15, 1988|

Fibrin-Fibronectin Compounds in Human Ovarian Tumor Ascites and TheirPossible Relation to the Tumor Stroma1

Olaf Wilhelm, Reimar Hafter, Eva Coppenrath, MaryAnn Pflanz, Manfred Schmitt, Rudolf Babic, Reinhold Linke,Wolfgang Gössner,and Henner Graeff2

Frauenklinik [O. W., R. H., M. P., M. S., H. G.] und Institut fürAllgemeine Pathologie und Pathologische Anatomie der Technischen Universität[E. C., R, B., W. G.],München;Institut fürImmunologie [R. L.] der Ludwig-Maximilians-Universität, Munich, Federal Republic of Germany

ABSTRACT

Covalenti) linked heterogeneous fibrin-fibronectin compounds weredetected in ascitic fluid of 31 patients with advanced ovarian cystadeno-carcinoma by means of enzyme-linked immunosorbent assay techniques,immunoaffinity chromatography, and Western blot analysis. Depositionof fibrin and fibronectin could also be demonstrated ¡mmunohistochemi-cally in Carnoy-fixed tissue sections. Fibrin and fibronectin were foundin the tumor stroma within tumor nests and more prominently in stromasurrounding the tumor nests. The association of fibrin and fibronectinwas especially pronounced in the stroma surrounding the tumor islands.Fibronectin was also found to be associated with stroma cells. Areaswithin the tumor stroma showed superimposed staining for both fibrinand fibronectin supporting the assumption that the covalenti)' linkedfibrin-fibronectin conjugates found in ascitic fluid may stem from theprovisional tumor stroma by proteolytic release.

INTRODUCTION

Advanced ovarian carcinomas grow mainly by spreading onthe surface of the abdominal cavity. Ascitic fluid is producedearly and is in FIGO3 IH/IV stages [Stage III: growth involving

one or both ovaries with i.p. métastasesincluding small bowelor omentum; Stage IV: distant or parenchymal liver métastases(1)] in constant contact with the tumor cells. Growth of humanovarian tumors may be related to the formation of a provisionalfibrin-fibronectin gel matrix surrounding the tumor (2, 3). Thisprovisional tumor stroma is replaced by the mature tumorstroma mainly consisting of reactive connective tissue and thecells also responsible for vascularization of the tumor stroma.

Serine proteases such as plasmin and the u-PA are known todegrade the provisional tumor stroma and release degradationproducts of fibrin and fibronectin into ascitic fluid of ovariancancer patients (4-6). The presence of such degradation products in ascitic fluid can be demonstrated by sensitive ELISAtechniques and high performance liquid chromatography including /V-terminal amino acid analysis (7, 8). Increase of theproteolytic fibrin split product D-dimer has served as a sensitiveindicator when monitoring the extent and course of the disease(6, 7). We now present evidence that cross-linked fibrin-fibronectin conjugates found in ascitic fluid of ovarian cancer patients may also stem in part from the provisional tumor stroma.

Received 8/18/87; revised 2/29/88; accepted 3/14/88.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by the Deutsche Forschungsgemeinschaft, Bonn-Bad Godesberg, FRG (Sonderforschungsbereich 207, A-2, Munich).

2To whom requests for reprints should be addressed, at Frauenklinik derTechnischen UniversitätMünchen.Ismaningerstrasse 22, D-8000 Munich, WestGermany.

3The abbreviations used are: APAAP. alkaline phosphatase anti-alkaline phos-phatase complex; BSA, bovine serum albumin; D-dimer, plasmin-derived terminaldegradation product of fibrin covalently linked through y-t chains; ELISA,enzyme-linked immunosorbent assay; F-CB3, fragment of fibrinogen a chainderived by cyanogen bromide cleavage; u-PA, urokinase plasminogen activator;FIGO, FédérationInternationale de Gynécologieet d'Obstétrique(1); Ig, immu-noglobulin; HRP, horseradish peroxidase; PAP, peroxidase anti-peroxidase complex; PBS, phosphate-buffered saline; PBST, PBS + 0.05% Tween 20; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis.

MATERIALS

The suppliers of the following reagents and antibodies are indicatedin parentheses.

Reagents. Levamisole, Tween 20, o-tolidine dihydrochloride, avidin-HRP (Sigma Chemicals Co., St. Louis, MO); 4-chloro-l-naphthol(Merck, Darmstadt, FRG); BSA (Behring, Marburg, FRG); fibrinogen(Deutsche Kabi, Munich, FRG); diaminobenzidine, naphthol-AS-MX-phosphate, Fast Blue BB, Fast Red TR (Serva, Heidelberg, FRG);nitrocellulose sheet BA85 (Schleicher & Schüll,Dassel, FRG); acryl-amide (>99.9% pure), urea (BIO-RAD, Munich, FRG); Roti-Histol

(Roth, Karlsruhe, FRG).Monoclonal Antibodies. Mouse anti-human fibronectin (IMCO,

Stockholm, Sweden); HRP-conjugated mouse anti-human fibrinogen (DD-4D2/182) and mouse anti-human D-dimer (DD-3B6/22)(MAMO, Brisbane, Australia). The production of mouse anti-humanF-CB3, detecting an epitope within the alpha chain sequence A«241-476 of human fibrinogen, was described elsewhere (9).

Polyclonal Antibodies. Ig rabbit anti-human fibrinogen, Ig rabbit anti-human fibronectin. Ig swine anti-rabbit Ig, Ig rabbit anti-human fibronectin (HRP-conjugated), PAP-rabbit, APAAP-mouse (Dakopatts,Hamburg, FRG); IgG sheep anti-mouse IgG (HRP-conjugated), IgGgoat anti-rabbit IgG (HRP-conjugated), IgG rabbit anti-mouse IgG(HRP-conjugated), IgG goat anti-rabbit IgG, F(ab')2 goat anti-rabbit

IgG (biotin conjugated), IgG rabbit anti-mouse IgG (biotin-conjugated)(Sigma Chemical Co., St. Louis, MO).

METHODS

Immunoelectrotransfer (Western Blot)

SDS-PAGE was performed according to Lacrimili (10) using aseparation gel of 4% acrylamide and a stacking gel of 3% acrylamide.Samples were prepared for electrophoresis by mixing an aliquot ofsample with one aliquot of sample buffer (4 ml 10% SDS (v/v), 2 ml0.5 M Tris-HCl, pH 6.8, 4 g urea) at 100°C(10 min). After electropho

resis, proteins were transfered to nitrocellulose sheets according toTowbin et at. (11) at 60 V (8 h, 10°C).The nitrocellulose sheets wereincubated with 2% BSA-Tris-buffered saline (12 h, 4'C) and then

reacted with a 1:200 dilution in PBST of IgG rabbit antifibrinogen (16h, 4"C) or a 1:30 dilution of tissue culture supernatant of monoclonalantibody to F-CB3, respectively (3 h, 37'C). After washing in PBST,

the nitrocellulose sheets were reacted with biotinylated anti-rabbit IgGor biotinylated anti-mouse IgG, respectively (1:200 dilution in PBST,3 h, 25°C).After washing in PBST the sheets were reacted with avidin-HRP (1:200 in PBST, 3 h, 25°C),washed in PBST and then peroxidase

activity revealed with a substrate solution consisting of 25 mg 4-chloro-l-naphthol dissolved in 10 ml 60% (v/v) cold ethanol containing0.006% H2O2.

ELISA

ELISAs were performed in 96-well fiat-bottomed polystyrene plates(Immunition. Dynatech, Denkendorf, FRG). Volumes: 100 «il/well.Before the addition of antigen the plates were coated with 200 ^I/wellof 2% BSA in PBS (2 h, 37°C).After binding of the various antibodies

the peroxidase reaction was started by the addition of 100 u\ of 2.65mM tolidine in 0.05 Mcitrate buffer (pH 4.0), containing 0.003% H2O2.After 10 min (23"C) the reaction was stopped by the addition of 50 ul

1 M H2SO4. The yellow reaction product was determined photometri-

3507

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

cally at 450 nm in a 96-well microtiter plate reader (Titertek Multiskan,Flow Laboratories, Bonn, FRG).

Measurement of D-dimer and Other Cross-Linked Fibrin DegradationProducts Containing the y-y Configuration. A two-step sandwich technique essentially according to the MAbCO protocol (Dimertest) wasused with monospecific monoclonal antibody DD-3B6/22. Captureantibody [DD-3B6/22 (PBS, 2 Mg/ml, 16 h, 23'C)]; antigen dilution

[plasma (1:5), ascine fluid (1:1000), peritoneal exúdate (1:20 and1:100), incubation (1 h, 23°C)];tagging antibody [HRP-conjugatedpanspecific monoclonal antibody DD-4D2/182 (1:8000, 1 h, 23'C)];standard curve [15.6-1000 ng D-dimer standard provided by MAbCO].All dilutions and washes were performed with PBST. This test isdescribed in more detail elsewhere (7, 12, 13).

Measurement of Fibrin(ogen). Capture antibody [IgG rabbit anti-human fibrin(ogen) (2 Mg/ml, 0.1 M bicarbonate/carbonate buffer, pH9.6, 16 h, 23°C)];antigen dilution [plasma (1:5000), ascitic fluid

(1:1000), peritoneal exúdales (1:1000 and 1:5000)]; tagging antibody[HRP-conjugated panspecific monoclonal antibody DD-4D2/182(1:8000, 1 h, 23°C)];standard curve [15.6-1000 ng/ml of fibrinogen in

human standard plasma (Behring, Marburg, FRG)]. All dilutions andwashes were performed with PBST.

Measurement of Fibronectin. Capture antibody [IgG rabbit anti-human fibroin-dill (2 Mg/ml, 0.1 M bicarbonate/carbonate buffer, pH9.6,16 h, 23°C)];antigen dilution [plasma (1:1000), ascitic fluid (1:500),

peritoneal exúdales (1:500 and 1:100)]; tagging antibody [HRP-conjugated IgG rabbit antifibronectin (1:1000, 1 h, 23'C)]; standard curve[15.6-1000 ng/ml of fibronectin in human standard plasma]. All dilutions and washes were performed with PBST.

Determination of Fibrin-Fibronectin Compounds. A two-way approachwas performed. First, microtiter plates were coated with antibodies tofibronectin and then the fibrin-fibronectin compounds detected by anantibody to fibrin(ogen). The opposite approach was performed bycoating the microtiter plates with an antibody to fibronectin and thendetection of the fibrin-fibronectin compounds by antibodies to fi-brin(ogen):

(a) First approach: Capture antibody [IgG rabbit anti-human fi-brin(ogen) 2 Mg/ml, 0.1 M bicarbonate/carbonate buffer, pH 9.6, 16 h,23'C)]; antigen dilution [plasma (1:100), ascitic fluid (1:100), peritoneal

exúdales (1:100)]; tagging antibody [HRP-conjugated IgG rabbit antifibronectin (1:1000, 1 h, 23°C)];standard curve [15.6-1000 ng/ml of

fibronectin in human standard plasma. The standard curve was takenfrom a parallel run of a fibronectin ELISA].

(b) Second approach: Capture antibody [IgG rabbit anti-humanfibronectin (2 Mg/ml, 0.1 M bicarbonate/carbonate buffer, pH 9.6, 16h, 23°C)];antigen dilution [plasma (1:100), ascitic fluid (1:100), peri

toneal exudates (1:100)]; tagging antibody [HRP-conjugated monoclonal antibody DD 4D2/182 (1:8000, 1 h, 23'C)]; standard curve[15.6-1000 ng/ml of fibrinogen in standard human plasma. The standard curve was taken from a parallel run of a fibrinogen ELISA). Alldilutions and washes were performed with PBST.

Miscellaneous

Clottable fibrinogen was determined according to Hafter and Graeff(14).

Immunoadsorption

2 g of BrCN-activated Sepharose 4B (Pharmacia, Freiburg, FRG)was swollen and washed according to the manufacturer's recommen

dations. The gel was resuspended in 8 ml 0.1 M NaHCO3 (pH 8.0),containing 2 mg of Ig rabbit antifibronectin (3 h, 20*C and then 16 h,4'C). Excess binding sites were blocked by resuspending the antifibro

nectin gel in 0.1 M Tris-HCI (pH 8.0). The gel was washed in 0.1 Mcitrate buffer (pH 4.0), resuspended in 10 ml Tris-HCI (pH 8.0), andthen 2 ml of this gel suspension was reacted with 500 M'of ascitic fluid(16 h, 4°C,and then 1 h. 20°C).The gel was washed three times with

10 ml of 0.1 M Tris-HCI (pH 8.0), containing l M NaCI. Desorptionof gel-bound material was achieved by incubation of the gel in 1 ml of0.1 M Tris-HCI -l- l M NaCI, saturated with urea. More than 98% ofthe IgG antifibronectin was covalently bound to the gel matrix.

Sample Preparation

Plasma Samples. Blood was collected from patients with advancedovarian cancer and healthy female donors as controls by venipunctureand anticoagulated by the addition of nine parts of blood to one partof anticoagulant [0.13 M sodium citrate, 0.06 M Tris, 0.05 M Na2EDTAand 2000 KIT aprotinin/ml (pH 7.4)). Plasma was obtained by centrifugation (10 min, 3000 x g, then 5 min, 5000 x g).

Ascitic Fluid. Ascitic fluid was aspirated by percutaneous punctureor during laparotomy from patients with advanced ovarian cancer(FIGO III/IV). Anticoagulant (1:10 by volume) was added. The sampleswere centrifuged immediately as described above for blood.

Peritoneal Exúdate.Peritoneal exudates were obtained under surgeryfrom patients (e.g., myoma of the uterus, cesarean section, and pelvicinflammatory disease) and processed as described for the collection ofascitic fluid. Samples of plasma, ascitic fluid, and peritoneal exúdatewere snap frozen in liquid nitrogen and stored at -20°C.

Tissue Preparation

Specimens of 17 ovarian serous cystadenocarcinomas (11 tumorsG3, six tumors G2) and two undifferentiated carcinomas of the ovarywere investigated. The tissues were fixed in Carnoy solution (15) andembedded in paraffin. Sections were cut and mounted on microscopeslides. These paraffin-cm bedded sections were dewaxed in Rot i-I fistol(a xylene substitute), followed by isopropanol and 96% (v/v) ethanol,and then rehydrated in PBS.

Immunohistochemical Detection of Fibrin(ogen) and Fibronectin on Tissue Sections

Dewaxed paraffin-embedded tissue sections of Carnoy-fixed ovariancarcinomas were processed for the detection of flbrin(ogen) and fibronectin. First, the presence of fibrin(ogen) or fractions thereof, wasrevealed by use of the peroxidase-antiperoxidase staining method (16).Endogenous peroxidase activity was inhibited by the action of 7.5%H2O2 (5 min, 20°C).Subsequently, the tissue sections were treated with2.3% periodic acid (5 min, 20'C), and then washed with PBS. The

tissue sections were covered with 20% (v/v) normal swine serum (20min, 20°C)which was replaced by a 1:100 dilution of Ig antifibrin(ogen)in 20% normal swine serum (20 min, 20°C).After washing in PBS a

1:50 dilution of Ig swine anti-rabbit Ig in PBS was added (20 min,20'C). After washing in PBS a 1:100 dilution of rabbit-PAP in PBSserum was added (20 min, 20°C).After washing in PBS the peroxidase-

dependent staining of tissue sections was revealed using 0.6 mg/mldiaminobenzidine in 0.05 M Tris-HCI, pH 7.6 (brown color) supplemented with 0.015% H2O2.

The slides were rinsed in PBS and processed for the presence offibronectin or fractions thereof by use of the alkaline phosphatase anti-alkaline phosphatase (APAAP) staining method (17). Subsequently, a1:50 dilution of monoclonal antibody to fibronectin was added in 20%normal rabbit serum (60 min, 20"C). After washing in PBS the tissue

sections were reacted with a 1:50 dilution of Ig rabbit anti-mouse Ig inPBS (30 min, 20'C), washed with PBS and then a 1:50 dilution ofmouse-APAAP in PBS was added (30 min, 20"C). After washing in

PBS the alkaline phosphatase-dependent staining was developed by 0.2mg/ml Naphthol-AS-MX phosphate in combination with 10 mg/ml?Fast Blue BB (blue color) or Fast Red TR (red color) in 0.2 M Tris-HCI (pH 8.5) containing 1 HIMlevamisole to block intrinsic alkalinephosphatase activity (20 min, 20°C).The tissue sections were washed

in PBS and water and mounted in glycerol-gelatine. Controls wereperformed by omitting the first antibody or by substituting the firstantibody by irrelevant IgG-antibodies of the relevant species.

RESULTS

Detection of Fibrin-Fibronectin Compounds in Ascitic Fluidfrom Patients with Advanced Ovarian Cystadenocarcinoma byELISA. Ascitic fluid (n —31) and corresponding plasma (n =24) of patients with advanced ovarian cystadenocarcinomaswere investigated by ELISA techniques for the presence of

3508

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

fibronectin- and fibrin-related antigen, cross-linked fibrin degradation product (D-dimer), clottable fibrinogen, and total protein. Peritoneal exúdales of a group of patients with benigngynecological diseases (n = 15) and plasma («= 10) obtainedfrom healthy donors served as controls. Significantly increasedlevels of the fibrin degradation product D-dimer were detectedin ascitic fluid from cancer patients compared to controls (Table1).

The presence of fibrin-fibronectin compounds (or degradation products thereof) in the patients' ascitic fluid and plasma

was also demonstrated (Fig. 1 and Table 1). Two differentapproaches by ELISA techniques were employed to demonstrate the presence of fibrin-fibronectin compounds in humanbody fluids. In the first approach, complexes were immobilizedon microtiter plates by antibodies to fibrin(ogen) and thendetected by antifibronectin. In the opposite approach thesecomplexes were immobilized on microtiter plates by use ofantibodies to fibronectin and then detected by anti-fibrin(ogen).Compared to the total amount of fibronectin (232 Mg/m') andfibrin(ogen) (644 Mg/ml) in ascitic fluid detected by ELISA, thecontent of fibrin-fibronectin compounds in ascitic fluid (7.4 pg/ml) or patient's plasma (3.7 Mg/ml) (reacting with both, anti

bodies directed to fibrin(ogen) and fibronectin) was low. Asciticfluid: 1.15% (w/w) of total fibrin(ogen); 3.75% of total fibronectin. Plasma: 0.1% of total fibrin(ogen); 1.04% of total fibronectin. Fibrin-fibronectin compounds were also detected inperitoneal exúdales and in plasma of healthy volunteers. Theamounts, however, were very low. Peritoneal exudates: 0.17%of total fibrin(ogen); 2.34% of total fibronectin. Control plasma:0.12% of total fibrin(ogen); 1.1% of total fibronectin. Theconcentrations of the fibrin-fibronectin compounds in asciticfluid, peritoneal exudates, and plasma are depicted in Fig. 1/4.The relative amounts of fibronectin and fibrin(ogen) measuredin the fibrin-fibronectin compounds were also plotted (Fig. IB).Evidently, an approximate 1:1 ratio on a weight basis of fibronectin antigens to fibrin antigens in fibrin-fibronectin compounds in ascitic fluids and peritoneal exudates is found.

Isolation of Covalenti) Linked Fibrin-Fibronectin Compoundsfrom Ascitic Fluid and Identification by Immunoelectrotransfer(Western Blot). Fibrin-fibronectin compounds from asciticfluids of three patients with advanced ovarian cystadenocarci-noma were isolated by incubating the ascitic fluid with anti-fibronectin-Sepharose 4B (for details see "Materials and Methods"). Rabbit Ig anti-human fibronectin proved to be specific

for intact fibronectin and its degradation products and did notbind to fibrin(ogen) in Western blot analysis (data not shown)which allowed the specific isolation of fibronectin and fibronectin conjugates from ascitic fluid. No unspecific binding of fibrin-fibronectin compounds to the Sepharose matrix was detectedwhen substituting rabbit IgG anti-human fibronectin by rabbitIgG anti-bovine serum albumin. The isolated fibronectin-con-taining material was subjected to 4% SDS-PAGE followed byelectrotransfer to nitrocellulose sheet. Fibrin-fibronectin conjugates were detected by staining the blots with Ig rabbit anti-human fibrin(ogen) (Fig. 2, lanes a-c). Ig rabbit anti-humanfibrin(ogen) did not react with fibronectin determinants inWestern blot analysis (data not shown).

The stained bands most probably represent fibrin-fibronectinconjugates as they proved to be stable after urea and SDS-treatment at 100°C.Up to seven bands covering a range of M,

50,000-500,000 were detected in the ascitic fluid of all threeovarian cancer patients. Polypeptide chains with a molecularweight of <250,000 showed a high degree of homology in allthree tumor patients whereas different polypeptide chains withhigher molecular weights were observed in the three differentascitic fluids.

Further information on the nature of these compounds wasachieved by reacting the blots with a monoclonal antibodyspecifically directed to a determinant within the a chain sequence Aa24i-476of fibrin (F-CB3) (Fig. 2, lanes d and e). Twoof the seven bands with an apparent molecular weight of110,000 and 500,000 detected in fibrin-fibronectin conjugatesby anti-fibrin(ogen) (Fig. 2, lanes a-c) also reacted with amonoclonal antibody to F-CB3 indicating that these two cova-

Table 1 Quantitation offibrin(ogen), flbronectin, fibrin-fibronectin compounds, and the fibrin degradation product D-dimer in plasma,ovarian cancer ascitic fluid, and peritoneal exudates

Data are mean values ±SD and expressed as Mg/ml and also ¡nmolar terms. The range of the values is also indicated. The calculation of pmol/ml is based on thefollowing molecular weights: fibrinogen, M, 340,000; fibronectin, M, 440,000; D-dimer, M, 186,000. FN, fibronectin; Fi, fibrin(ogen).

Tumor ascites (n =31)Mg/mlRangepmol/mlRangePeritoneal

exúdate(n =IS)Mg/mlRangepmol/mlRangePlasma

(n = 24)(patient)Mg/mlRangepmol/mlRangePlasma

(n = 10)(control)Mg/mlRangepmol/mlRangeFi-FN

compounds"7.4

±7.23-3017

±167-682.1

±0.71.0-3.44.8

±1.62.3-7.73.7

±0.82.5-5.28.4

±1.85.7-11.82.8

±0.62.6-3.66.4

±1.45.9-8.2FN-Fi

compounds'8.7

±7.03-3025

±209-882.3

±1.20.1-3.86.8

±3.50.3-11.24.5

±1.92-8.513

+5.66-253.0

±0.32.4-3.88.8

±0.97.0-11.2Fibronectin

antigen232

±13460-610527

±304136-1,38698

±9630-330223

+21868-750432

±140265-610982

±318375-1,386263

±51183-343598

±116414-780D-dimer255

±17952-7801,370

±960280-4,2004.3

±3.31.0-1123

±185-592.2

±1.90.2-3.96.4

±5.50.6-11.50.1

±0.040.1-0.20.5

±0.20.3-1.0Fibrinogen

antigen644

±29370-1,0501,894

±861205-3,0881,290±

1,53040-5,0003,790

±4,500118-14,7003,695

±1,0651,000-5,60010,870

±3,130294-16,4702,325

±9021,500-5,7506,838

±2,6524,412-13,970Fibrinogen

clottable57

±2220-91167

±6060-270985

±1,67020-4,6002,897

±4,91258-13,5303,650

±1,0201,000-5,60010,750

±3,000294-16,4702,350

±801,500-4,7006,912

±2354,412-13,824Total

protein42,800

±10,00020,000-58,50045,500

±14,5801,600-47,20061,

100±7,50046,500-79,00061,920

±3,19057,200-66,800

" Absorbed to microtiter plates by antifibrin(ogen) and detected by antifibronectin.* Absorbed to microtiter plates by antifibronectin and detected by antifibrin(ogen).

3509

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

30-

25

20

IS

10

viT r

?

v

..II..

FNFiTUMORASCITESn.

3lFN

FiPERITONEALEXUDATESn,

15FN

FiPLASMAPATIENTSn.

24FN

FiPLASMACONTROLSn,

10

30

20

IIIOO

15.

u. 10

ASCITES (n, 31)

OVARIANCANCER "B

PERITONEAL (n = 15)EXUDATES /

10 16 20 25 30

FIBRONECTIN < ufl / ml >

Fig. 1. .I. distribution of concentrations of fibrin(ogen) (fi) and fibronectin(FN) in fibrin fibremi-ciin compounds determined in ovarian tumor asciti-v peritoneal exúdales, and plasma of patients and healthy donors. Fibrin-fibronectincompounds in ascitic fluid were significantly elevated compared to values obtainedwith peritoneal exúdales (p < 0.001). Statistical evaluation was performed usingthe Mann-Whitney test for independent random tests. B, fibrin(ogen)-fibronectindiagram for concentrations of fibrin(ogen) and fibronectin determined in compounds in ascitic fluid and peritoneal exúdales by plotting the amounts offibrin(ogen) versus flbronectin. •.Single data; O, mean value of data. Correlationcoefficient, 0.574.

lently linked fibrin-fibronectin compounds contain this part ofthe a chain of cross-linked fibrin.

Evaluation of Fibrin and Fibronectin Deposits in Fixed TissueSections of Advanced Ovarian Cystadenocarcinomas by Immu-nohistochemistry. Tissue sections of 19 different cystadenocar-

cinomas of the ovary were examined immunohistochemicallyfor the presence of fibrin, fibronectin, and degradation productsthereof. Tissues were fixed in Carnoy solution and then subsequently processed to detect the various antigens as described in"Materials and Methods." In all 19 tissues we stained antigens

reactive with antibodies to fibrin and fibronectin, respectively.In Fig. 3-6, examples of various regions in two different cys-tadenocarcinomas of the ovary are shown. The major findingis, that in addition to regions which clearly reacted with antibodies against fibrin or fibronectin, there are also parts withinthe tumor stroma in which superimposed staining of both, fibrinand fibronectin, can be demonstrated. Superimposed stainingmay indicate the close association of fibrin and fibronectin inthe provisional tumor stroma.

The majority of antigens reacting with antibodies directed tofibrin or fibronectin are located in the tumor stroma surrounding the tumor nests. This was less pronounced in the stromawithin tumor nests. In addition, obvious association of fibronectin with stroma cells can be demonstrated (Fig. 6).

Regions within the tumor stroma reacting with antibodies tofibrin also reacted with antibodies to the fibrin a chain determinant Att24i-476and the D-dimer epitope in cross-linked fibrin(data not shown). These observations support the notion thatfibrin degradation products found in ascitic fluid may be derivedfrom the provisional tumor stroma by proteolytic degradation.

DISCUSSION

The presence of covalently linked fibrin-fibronectin conjugates was demonstrated in ascitic fluid from patients withadvanced ovarian carcinoma. Fibrin and fibronectin, the majorconstituents of the provisional tumor stroma were also foundin tumor tissue of these patients by immunohistochemistry. Theprovisional tumor stroma serves as a matrix for new capillariesand the immigration of macrophages and fibroblasts. It isproteolytically degraded and subsequently replaced by matureconnective tissue stroma (2,18). In advanced ovarian carcinomathe degradation products which are released from the fibrin-fibronectin gel by proteolysis are drained into the abdominalcavity and are components of the tumor ascites. Access of thecomponents of ascitic fluid to the blood via lymphatic drainageis supported by the finding of high levels of cross-linked fibrindegradation products such as D-dimer in ascitic fluid and bloodin ovarian cancer patients (4-7, 19, 20). In the present study inaddition to D-dimer products we also find elevated levels offibrin-fibronectin compounds in blood and ascitic fluid of patients with advanced ovarian carcinomas.

The proteolysis of the provisional matrix is mainly regulatedby the serine protease u-PA (22, reviewed in Ref. 21). Thepresence of this enzyme in ovarian carcinoma was alreadydemonstrated in 1976 by Astedt and Holmberg (23). Immunocytochemically it was found in Lewis lung carcinoma to belocated in regions of invasive tumor growth (24). u-PA intumors may be produced by tumor and/or stroma cells (21, 25).Our own observations located u-PA in ovarian carcinoma to beassociated with tumor cells and in the tumor stroma.4 By

activation of plasminogen to plasmin, degradation of fibrin andfibronectin is initiated (20, 26). It was also reported that limiteddegradation of fibronectin by u-PA might occur without plasmin activation (27). Compared to nonmalignant peritonealexudates we found elevated amounts of cross-linked fibrindegradation products [D-dimer and higher molecular weight

Jänicke,Schmitt, and Babic, unpublished data.

3510

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

0)'0

Co

is

500

340310240

145110

50 —

TOP

*

Fig. 2. Immunoblot analysis of covalently linked fibrin-fibronectin antigens isolated by antifibronectin and recognized by antibodies to nbrin(ogen) or the fibrin achain determinant F-CB3. Ascitic fluid of three different patients was absorbed to matrix-bound antifibronectin and then desorbed with urea (for details see "Materialsand Methods"). The desorbed antigens were subjected to SDS-PAGE (4% acrylamide) under nonreducing conditions. Antigens separated by gel electrophoresis weretransferred electrophoretically to nitrocellulose paper and then stained with antifibrinogen (lanes a-c) or anti-F-CB3 (lanes d and e). Lanes a-c, patients 1-3; lanes dand e, patients 1 and 2.

Fig. 3. Poorly differentiated cystadenocar-cinoma (G3) of a patient with FIGO III. Areaselected represents surface of the tumor. Im-munohistochemical staining for the presenceof fibrin (brown) and fibronectin (blue). Fibrinand fibronectin are located in the tumorstroma. Microscopic magnification, x 40.

products XDP (28)] in ascitic fluid of ovarian cancer patients(5, Table 1). This is also true for fibrin-fibronectin compounds(Table 1). Elevated levels of fibrin-fibronectin compounds werealso found in corresponding plasma samples of these patients.Small amounts of these compounds were present in plasma ofhealthy donors. Fibrin can interact with fibronectin either bynoncovalent binding (complex formation) or by factor XIHa-transglutaminase-induced covalent binding (26, 29-31).

To analyze the nature of interaction of the fibrin-fibronectincompounds found in ascitic fluid, immunoabsorption experiments with antifibronectin immobilized to Sepharose were performed. The desorbed antigens were subjected to SDS-PAGEand the covalent binding of fibrin to fibronectin in the conjugates was revealed using antifibrin(ogen) as the detecting antibody. Ascitic fluid from three different patients was examined.Differences in the molecular weights of polypeptide chains with

molecular weights of >250,000 were evident but also corresponding chains were found. These covalently linked fibrin-fibronectin conjugates most probably do not contain intactfibrin a chains since the C-terminal part of the a chain isdegraded early (32). A monoclonal antibody directed to the achain determinant F-CB3 reacted only with epitopes within twopolypeptide chains of approximate molecular weights of110,000 and 500,000. Obviously, in the other conjugates, fibronectin antigen is linked to a chain remnants which do notcontain epitopes of the a chain sequence Aa24i-»76which wouldotherwise be detected by anti-F-CB3 (9, 33). Evidently, most ofthe low molecular weight polypeptide chains reacting with bothantifibronectin and antifibrin(ogen) do only contain remnantsof fibrin(ogen) and fibronectin.

The origin of fibrin-fibronectin conjugates in ascitic fluid isstill a matter of debate. According to Henderson et al. (34)

3511

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

Fig. 4. Same tumor and immunohisto-chemical staining as in Fig. 3 at higher magnification. Area selected represents central partof the tumor. Fibronectin (blue) forms a reticular network between tumor cells. Fibrin(brawn) is mainly located in the stroma surrounding the tumor nests. Microscopic magnifìcation,x 400.

Fig. 5. Higher magnification of surface-near portion of the same tumor as in Fig. 3.Superimposed brown and blue colors demonstrate the close association of fibrin and fibro-nectin in the provisional tumor stroma andmay represent the same covalently linked fi-brin-fibronectin conjugates as detected in thepatients' ascitic fluid by Western blot analysis.

Microscopic magnification, x 400.

plasma/ascitic fluid exchange and enzymatic effects can beevaluated by plotting the ratios of concentrations of proteinsdetermined in plasma and ascitic fluid versus the natural logarithm of the molecular weights of the respective proteins inconsideration. Depletion of fibrinogen in ovarian cancer asciticfluid and enrichment of its degradation products indicate thatthis protein was metabolized extravasally (6). Ascitic fluid is inconstant contact with tumor cells in advanced carcinoma.

Fibrin and fibronectin were demonstrated immunohisto-chemically in the tumor stroma within tumor nests and morepronounced in the stroma surrounding the tumor nests. Particularly, the association of fibrin and fibronectin was more prominent in the stroma surrounding the tumor islands. Part of thefibronectin visualized in tumors may be derived from stromacells, such as fibroblasts (35, 36). We also observed in ovarian

cancer tumor stroma fibronectin-positive cells. This cellularsource of fibronectin might be important in tumor stromageneration. Webb and Lin (37) evaluated urinary fibronectinlevels as a biomarker of prostatic cancer assuming proteo!)ticrelease of fibronectin antigen from the tumor site. Fibrin identified in tumor tissue by antibodies to an epitope in the cross-linking region (D-dimer) most probably stems from extrava-sated fibrinogen. The soluble D-dimer containing cross-linkeddegradation products of fibrin are enriched in ascitic fluid. Itmight well be assumed that these components stem from thedegraded fibrin in tumor stroma. In parallel, in all the tumorsinvestigated we noticed areas of superimposed staining for bothfibrin and fibronectin in stroma surrounding tumor nests. Mostprobably these fibrin-fibronectin associates are the source ofthe fibrin-fibronectin compounds found in ascitic fluid after

3512

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

Fig. 6. Poorly differentiated cystadenocar-t inuma (G3) of the ovary (FIGO III). Differenttumor than in Figs. 3-5. Immunohistochemi-cal staining of fibrin (Immn) and fibronectin(red). Nuclei (blue) stained with hematoxylin.H. enlargement of the area indicated (rectangle) in A. Fibrin and fibronectin are located inthe tumor stroma surrounding the tumor nests.There is a close association of fibrin and fibronectin in the tumor stroma (black arrows).Fibronectin is also associated with stroma cells(white arrows). Microscopic magnification ofA, x 160. Microscopic magnification of B,x 1000.

»V*

proteolysis. However, at the light microscopic level it is difficultto decide if these areas of superimposed staining representcovalently linked fibrin-fibronectin conjugates.

ACKNOWLEDGMENTS

The technical assistance of E. Sedlaczek, B. Obermeier, S. Kamp-kötter,J. Roth, B. Weinhold, and I. Hoepner is gratefully acknowledged. We thank MAbCO Limited for their generous gift of monoclonal antibodies to fíbrinogenand D-dimer.

REFERENCES1. Hermanek, P., and Sobin, L. H. (eds.), TNM Classification of Malignant

Tumours, 4., fully revised edition, International Union Against Cancer,Geneva. Berlin-New York: Springer Verlag, 1987.

2. Dvorak, H. F. Tumors: wounds that do not heal. New Engl. J. Med., 315:1650-1659, 1986.

3. Donati, M. B., and Semeraro, N. Fibrin-tumor cell interaction. In:G. Müller-Berghaus, U. Scheefers-Borchel, E. Selmayr, and A. Menschen (eds.), Fibrin-ogen and Its Derivatives, pp. 153-159. Amsterdam: Elsevier Science Publishers B. V.. 1986.

4. Graeff, H., and Halter, R. Detection and relevance of crosslinked fibrinderivatives in blood. Semin. Thromb. Haemost., í:57-68, 1982.

5. Hafter, R., Klaubert, W., Gollwitzer, R., von Hugo, R., and Graeff, H.Crosslinked fibrin derivatives and fibronectin in ascitic fluid from patientswith ovarian cancer compared to ascitic fluid in liver cirrhosis. Thromb. Res.,35: 53-64, 1984.

6. Robl, M., Jürgensmeyer,K., Hafter, R., Schröck,R., Wilhelm, O., Babic,R., and Graeff, H. Relation of fibrin and tumor-associated antigens to thespread of ovarian cancer. Fibrinolysis, /: 143-148, 1987.

7. Hafter, R., Schröck, R., von Hugo, R., and Graeff, H. Measurement ofcrosslinked fibrin derivatives in plasma and ascitic fluid with monoclonalantibodies against l> dinier using EIA and latex test. Scarni. J. Clin. Lab.Invest., «(Suppl. 178): 137-144, 1985.

8. Wilhelm, O., Henschen, A., Hafter, R., and Graeff, H. Tumor-associatedfibrinolysis in ovarian carcinoma—HPLC and .V terminal amino acid analysis reveal the pathway of degradation. Thromb. Haemost., 5*(1): 110, 1987.

3513

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

FIBRIN-FIBRONECTIN DEGRADATION IN OVARIAN CARCINOMA

9. Gollwitzer, R., Kulbe, G. E., Gabrijelcic, I ).. and Linke, R. P. A monoclonalantibody against the CNB-peptide F-CB3 of the A .»-chainof human fibrin-ogen. In: D. A. Lane, A. Henschen, and M. K. Jasani (eds.), Fibrinogen-Fibrin Formation and Fibrinolysis, Vol. 4, pp. 297-302, Berlin, New York:Walter de Gruyter, 1986.

10. Laemmli, I K. Cleavage of structural proteins during the assembly of headof bacteriophage T4. Nature (Lond.), 227: 680-685, 1970.

11. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteinsfrom polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Nati. Acad. Sci. USA, 76: 4350-4354, 1979.

12. Rylatt, D. B., Blake, A. S., Cottis, L. E., Messingham. D. A., Fletcher, W.A., Masci, P. P., Whitaker, A. N., Elms, M., Bunce, J., Webber, A. J., Wyatt,D., and Bundesen, P. G. An immunoassay for human D-dimer using monoclonal antibodies. Thromb. Res., 31: 767-778, 1983.

13. Whitaker, A. N., Elms, M. J., Masci, P. P., Bundesen, P. G., Rylatt, D. B.,Webber. A. J., and Bunce. I. Measurement of crosslinked fibrin derivativesin plasma. J. Clin. Pathol.. 37: 882-887, 1984.

14. Hafter. R., and Graeff, H. Estimation of soluble fibrin monomer complexesby agarose gel filtration. In: J. F. Davidson, M. M. Samama, and P. C.Desnoyers (eds.). Progress in Chemical Fibrinolysis and Thrombolysis, Vol.2. pp. 137-149. New York: Raven Press, 1976.

15. Babic, R. rinfili.! der Fixierung auf den immunohistochemischen Nachweisvon Tumormarkern und tumorassoziierten Antigenen. In: K. Hübner(Hrsg.),Verhandlungen der Deutschen Gesellschaft fürPathologie, 71. Tagung 1987,S. 423. Stuttgart-New York: Gustav Fischer, 1987.

16. Stemberger, L. A., Hardy, P. H., Cuculis, J. J., and Meyer, H. G. Theunlabelled antibody enzyme method of immunohistochemistry. Preparationand properties of soluble antigen-antibody complex (horseradish peroxidase-antihorseradish peroxidase) and its use in identification of spirochetes. J.Histochem. Cytochem., 18: 315-333, 1970.

17. Mason, D. Y.. and Summons. R. Alkaline phosphatase and peroxidase fordouble immunoenzymatic labelling of cellular constituents. J. Clin. Palhol.,}¡:454-460, 1978.

18. Dvorak, H. F. Abnormalities of hemostasis in malignancy. In: R. W. Cole-man, J. Hirsch. V. J. Marder, and E. W. Salzman (eds.), Hemostasis andThrombosis, pp. 1143-1157. Philadelphia: J. B. Lippincott, 1987.

19. Schrock, R., Hafter, R., Schmid, L.. Babic, R., Ulm. K., Gössner,W.. andGraeff, H. Tumorassoziierte Antigene und Fibrinderivate als Reaktionsprodukte des Ovarialkarzinoms. Geburtshilfe und Frauenheilkunde, 46: I-10, 1986.

20. Graeff, H., and Hafter, R. Clinical aspects of fibrinolysis. In: A. L. Bloom,and D. P. Thomas (eds.). Haemostasis and Thrombosis, pp. 245-254. Edinburgh: Churchill Livingstone, 1987.

21. Daño.K., Andreasen, P. A., Grondahl-Hansen, J., Kristensen, P., Nielsen,L. S., and Skriver, L. Plasminogen activators, tissue degradation and cancer.Adv. Cancer Res., 44: 139-266, 1985.

22. Layer, G. T., Cederholm-Williams. S. A., Gaffney, P. J., Houlbrook, S.,

Mahmoud, M., Pattison, M., and Burnand, K. G. Urokinase—the enzymeresponsible for invasion and metastasis in human breast carcinoma? Fibrinolysis, /: 237-240, 1987.

23. Astedt, B., and Holmberg. L. Immunological identity of urokinase andovarian carcinoma plasminogen activator released in tissue culture. Nature(Lond.), 262: 595-597, 1976.

24. Skriver, L., Larson, L. J., Kielberg, V., Nielsen, L. S., Andresen, P. B.,Kristensen, P., and Daño,K. Immunocytochemical localization of urokinasetype plasminogen activator in Lewis lung carcinoma. J. Cell. Biol., 99: 753-757, 1984.

25. Larsson, G., Larsson, A., and Astedt, B. Tissue plasminogen activator andurokinase in normal, dysplastic and cancerous squamous epithelium of theuterine cervix. Thromb. Haemost., 58(1): 822-826, 1987.

26. Hörmann,H. Interaction with fibrinogen and fibrin. In: J. McDonagh (ed.),Plasma Fibronectin, Structure and Function, pp. 99-120. New York: MarcelDekker, 1985.

27. Quigley, J. P., Gold, L. J., Schwimmer, R., and Sullivan, L. M. Limitedcleavage of cellular fibronectin by plasminogen activator purified from transformed cells. Proc. Nati. Acad. Sci. USA, 84: 2776-2780. 1987.

28. Francis, C. W., Alkjaersig, N., Galanakis, D. K., Graeff, H., Owen, J.,Gaffney, P., and Marder, V. J. Terminology for macromolecular plasmicderivatives of crosslinked fibrin. Thromb. Haemost., 57(1): 110-111, 1987.

29. Ruoslahti, £.,Hayman, E. G., Pierschbacher, M., and Engvall, E. Fibronectin: purification, immunochemical properties and biological activities. In: S.P. Colowick and N. O. Kaplan (eds.), Methods in Enzymology, Vol. 82, pp.803-831. New York: Academic Press, 1982.

30. Mosher, D. F. Fibronectin—relevance to hemostasis and thrombosis. In: R.W. Coleman, J. Hirsch, V. J. Marder, and E. W. Salzmann (eds.), Hemostasisand Thrombosis, pp. 210-218. Philadelphia: J. B. Lippincott, 1987.

31. Mosher, D. F., and Johnson, R. B. Specificity of fibronectin-fibrin crosslink-ing. Ann. NY. Acad. Sci., 408: 583-593, 1983.

32. Pizzo, S. V.. Schwartz, M. L., Hill, R. L., and McKee, P. A. The effect ofplasmin on the subunit structure of human fibrin. J. Biol. Chem., 248:4574-4583, 1973.

33. Gollwitzer, R., Hafter, R., Timpl, R., and Graeff, H. Immunological assayfor a carboxyterminal peptide of the fibrinogen «-chain in normal andpathological human sera. Thromb. Res.. //: 859-868, 1977.

34. Henderson, J. M., Stein, S. T., Kutner, M., Wiles, M. B., Ansley, J. D., andRudman, D. Analysis of twenty-three plasma proteins in ascites. The depletion of fibrinogen and plasminogen. Ann. Surg., 192: 738-742, 1980.

35. Stenman, S., and Vahen, A. Fibronectin in human solid tumors. Int. J.Cancer, 27: 427-435, 1981.

36. D'Ardenne, A. J., Burns, J., Sykes. B. C., and Benett M. K. Fibronectin and

type III collagen in epithelial neoplasms of gastrointestinal tract and salivarygland. J. Clin. Pathol., 36: 756-763, 1983.

37. Webb, K. S., and Lin, G. H. Urinary fibronectin. Potential as a biomarker inprostatic cancer. Invest. Urol., 17:401-404, 1980.

3514

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

1988;48:3507-3514. Cancer Res   Olaf Wilhelm, Reimar Hafter, Eva Coppenrath, et al.   Ascites and Their Possible Relation to the Tumor StromaFibrin-Fibronectin Compounds in Human Ovarian Tumor

  Updated version

  http://cancerres.aacrjournals.org/content/48/12/3507

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/48/12/3507To request permission to re-use all or part of this article, use this link

on March 28, 2021. © 1988 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from