Eur. J. Biochem. 195, 735-741 (1991) 0FEBS 1991 001429569100089X

A tetranectin-related protein is produced and deposited in extracellular matrix by human embryonal fibroblasts Inge CLEMMENSEN

’,Leif R. LUND’,

Lise CHRISTENSEN3 and Peter A. ANDREASEN4

’ Dakopatts Ltd and Department of Clinical Microbiology, Statens Serum Institute at Rigshospitalet, Copenhagen, Denmark ’ The Finsen Laboratory, Rigshospitalet, Copenhagen, Denmark Department of Pathology, Rigshospitalet, Copenhagen, Denmark Institute of Molecular Biology and Plant Physiology, University of k h u s , Denmark (Received March 19/August 22, 1990) - EJB 90 0310

Tetranectin is a tetrameric human plasma protein that binds to plasminogen kringle 4. Its amino acid sequence is homologous with the C-terminal parts of asialoglycoprotein receptors and proteoglycan core proteins. In the present study, we have demonstrated that the human embryonal fibroblast cell line WI-38 produce a tetranectinrelated molecule, which might, by several criteria, be similar to tetranectin from plasma. These criteria include immunoblotting analysis of conditioned cell medium revealing a protein band with MI 17,000, indistinguishable from the M , of plasma tetranectin. A preparation obtained by purification of conditioned medium by affinity chromatography on an anti-(plasma tetranectin) IgG column also contained the M , 17000 protein. This protein (partly purified from the conditioned medium) was shown by crossed immunoelectrophoresis to bind to heparin, CaClz and plasminogen kringle 4, as previously described for tetranectin in plasma. Importantly, this tetranectinrelated protein is not only present in conditioned culture medium, but the MI 17000 protein reacting with anti(plasma tetranectin) IgG was also present in the extracellular material, remaining after removal of WI-38 cells from the culture dishes, as demonstrated by immunoblotting analysis and immunocytochemical staining. We conclude that WI-38 cells produce a tetranectin-related protein and secrete it into the extracellular matrix. Tetranectin is a plasminogen-kringle-4-bindingprotein from plasma. It is a non-covalently bound tetrameric protein with four identical peptide chains [l]. Each peptide contains three disulphide bonds; no attached carbohydrate has been found [2]. Amino acid sequence homology has been found at the C-terminal part of the liver cell asialoglycoprotein receptors from man, rat and chicken, the C-terminal part of the core protein of a rat cartilage proteoglycan, a surfactant factor from canine lung and a lymphocyte receptor for the F, part of IgE [2 - 101. From these homologous sequences, it has been proposed that tetranectin may be the circulating C-terminal part of a receptor, as found in plasma for the F, part of IgE [2, lo]. The concentration of tetranectin in plasma varies with age and sex [ll], and has been found to be significantly decreased in patients with cancer, especially when disseminated 1121. The present study describes the secretion of a tetranectinrelated protein from human embryonic lung fibroblasts and its secretion into the extracellular matrix by these cells. These finding suggest that tetranectin may function as an anchor for plasminogen in the extracellular matrix. MATERIALS AND METHODS Materials

Lyophilized ammonium heparinate (1.5 x lo5 Ujmg) was from Leo Pharmaceuticals (Copenhagen, Denmark). AproCorrespondence to I. Clemmensen, Department of Clinical Microbiology, Statens Serum Institute at Rigshospitalet, Juliane Mariesvej 28’, DK-2100 Copenhagen, Denmark Abbreviation. HRP, horse radish peroxidase.

tinin (Trasylol) was from Bayer (Leverkusen, FRG). 6-Aminohexanoic acid was a generous gift from Kabi (Stockholm, Sweden). Agarose A-37 (Indubiose) was from Llndustries Biologigues Francaises (Gennevilliers, France). Sepharose 4B and DEAE-Sepharose CL 6B: was from Pharmacia (Uppsala, Sweden).

Cell culture The human embryonal lung fibroblast cell line WI-38 (ATCC CCL 75) was purchased from Flow Laboratories, Irvine, UK. The cells were seeded in 15-cm dishes and grown to confluence in 10% fetal calf serum. At confluence, which was attained in 2-4 days, each dish contained 5-6 x lo6 cells. The cultures were then washed with 0.01 M Na2HP04/ 0.15 M NaCl, pH 7.49 (NaCl/Pi), as described previously [13]. For immunocytochemical analyses cells were seeded in slide flasks (Nunc, Denmark). Cell number was calculated by determination of DNA [14]. Conditioned media and cells were harvested as described [13]. A few other cell lines were also tested for tetranectin production (see Results). Details are given in [131. Antibodies

Rabbit immunoglobulins against human plasma tetranectin (Dakopatts A/S, Glostrup, Denmark, code no. A 371 ; IgG concentration, 2.5 g/l) were purified on a tetranectinSepharose affinity column (6 mg tetranectin coupled to 5 ml Sepharose 4B) [15]. Bound antibody was eluted by 4 M KSCN and dialysed against 0.1 M Na2C03 (pH 8.0). Non-immune rabbit IgG for control in immunoblotting analysis was also

736 from Dakopatts (code X 903; IgG concentration 20.0 g/l). Biotinylation of affinity-purified anti-tetranectin IgG was performed by adding 0.4 mg biotinamidocaproate N-hydroxysuccinimide (Sigma, USA), dissolved in 8 pl dimethylformamide, to 3.75 ml(0.4 g/l) purified antibody. After incubation for 2 h at room temperature, the conjugated antibody was dialysed against 0.05 M Tris/HCl, 0.1 M NaCl (pH 7.2). This antibody was used in the ELISA for detection and quantification of tetranectin. For immunoblotting analysis, affinity purification of the commercial antibody was performed by elution from nitrocellulose paper. Briefly, 40 pg purified tetranectin was run in 10% SDS/PAGE. One lane of the gel slab was cut off and stained with Coomassie blue. The rest was transferred to 0.1 pm nitrocellulose paper (Schleicher and Schull, FRG) and incubated fro 1 h at room temperature in 50 ml 50 mg/l rabbit anti-tetranectin IgG. The strip corresponding to the position of tetranectin was cut out and incubated in 0.2 M glycine/HCI (pH 2.8) for 5 min. The paper was removed and the eluent was neutralized with 1.0 M Tris base. The concentration of the eluted antibody was 100mg/l as estimated by the absorbance at 280 nm. The antibody was used undiluted as primary antibody in immunoblotting analyses after addition of Tween 20 to a final concentration of 0.6%. Adsorption of anti-tetranectin antibodies was performed by passing the antibody through a column of tetranectin-linked Sepharose 4B. The resulting antibody was controlled for lack of tetranectin reactivity by immunoblotting.

column (100 ml Sepharose 4B). The column was washed with 0.05 M Tris/HCl, 0.6 M NaCl 0.6 M (pH 7.2) until the absorbance was below 0.010. Tetranectin bound to the column was eluted with 3.0 M KSCN. The purity of the protein prepared by kringle-4 chromatography and immunoaffinity chromatography was evaluated by SDSjPAGE: the concentration of the purified protein was measured by its absorbance at 280 nm [l]. Purijication of tetranectin from human plasma by conventional methods. This was performed by the following steps as recently described [I]. (a) Removal of plasminogen by lysineSepharose 4B; (b) (NH4)2S04 precipitation; (c) affinity chromatography on kringle-4 - Sepharose 4B; (d) ion-exchange chromatography; (e) gel chromatography, on Ultrogel AcA 34. Crossed immunoeletrophorrsisJ.si.~ This was performed in Trisibarbiturate buffer at pH 8.6. The second-dimension electrophoresis was performed in rabbit anti-(human tetranectin) IgG. Samples applied in the slit were purified tetranectin-immunoreactive material from WI-38 cell supernatants purified on anti-tetranectin-IgG Sepharose; as a control, tetranectin purified from human plasma was run on identical columns. The first-dimension electrophoresis was performed with or without 2 mM CaC12. First-dimension electrophoresis was also performed with the incorporation of 25000 U heparin into the 15-ml agarose gel.

Enzyme- link ed init??unosorhen t assay Tetranectin in WI-3X cell culture supernatants was determined by an ELISA a s recently described [I 11, but with biotinylated affinity-purified antibody at a concentration of 50 ng/l in place of horseradish-peroxidase (HRP)-conjugated anti-tetranectin IgG and with HRP-conjugated streptavidin (Dakopatts, code no. P 397), diluted 1 : 5000 as the detection system (Dakopatts, code no. P 397). Plates were read at 492 nm in a 2550 EIA reader (Bio-Rad, USA). Serial dilutions of purified tetranectin and pooled plasma were used as standards. Purification of >>roteins Purification 01'plasminogen kringle 4 and preparation of plasminogen-krin~le-4- Sepharose. This was performed as recently described [I]. 4pmol kringle 4 was linked to 8 g Sepharose. For affinity chromatography of concentrated WI38 conditioned medium on kringle-4-Sepharose 4B, the kringle-4 column was equilibrated with 0.05 M Tris/HCl, 0.1 M NaCl (pH 7.2). 10 ml 100-fold-concentrated conditioned medium, dialysed against the equilibration buffer, was applied to the column. The column was washed with the equilibration buffer until the absorbance to 280 nm was below 0.010. Adsorbed proteins were released by 1 mM 6-aminohexanoic acid. Flow-through and eluate were concentrated by ultrafiltration to 10 ml and analysed by immunoblotting. Tetranectin and tetraneetin-related protein. These were purified from human plasma or from concentrated conditioned WI-38 medium on rabbit anti-tetranectin IgG immobilized on Sepharose 4B (250 mg IgG coupled to 25 g Sepharose 4B). Prior to coupling to Sepharose 4B, the antibody was passed through a column of lysine-linked Sepharose 4B to remove rabbit plasminogen/plasmin present in rabbit IgG preparations [16]. 1 1 plasma or 50 ml concentrated conditioned medium, originating from 4 1, was applied to the

S D S/ PAGE and immunohlo tt ing SDSjPAGE was run in 10% polyacrylamide gel slabs. lmmunoblotting analysis of proteins transferred to 0. I-pm nitrocellulose sheets was performed by using affinity-purified anti-tetranectin IgG (100 ingil; see above), and alkaline-phosphatase-conjugated swine anti-(rabbit IgG) antibodies (Dakopatts code no. D 306; dilution of 1:500). As control, anti-tetranectin IgG was replaced with non-immune rabbit IgG (25 mg/l). Prestained SUSjPAGE standards (low range) were from Bio-Rad. Immunocytochemical staining o j WI-38 cells and extracellular mutrix WI-38 human embryonal lung fibroblasts were seeded on slide flasks and grown as described above. Some cells had been cultured in serum-free medium for 3 days. Conditioned medium was removed from dishes and replaced by NaC1/Pi. Some of the flasks were left for 36 h at 4°C. This procedure loosened the cells, which were subsequently released by applying a horizontal blow to the dishes. Total cell removal was assured by light microscopy. Slides with and without cells were fixed for 30 min at room temperature with 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.3. After washing with 0.1 M Tris/HCl, 0.15 M NaC1, 1%Triton X-100, pH 7.5 (buffer A), for 30 min, the preparations were exposed to 10% bovine serum albumin in 0.1 M Tris/HCl, 0.15 M NaC1, pH 7.5, for 30 min. The preparations were incubated with rabbit anti-(human tetranectin) IgG, 25 mg/l in 0.1 M Tris/ HCl, 0.15 M NaCl, 5% bovine serum albumin, pH 7.5 (buffer B), for 24 h at 4"C, followed by 24 h at room temperature. After washing with buffer A twice for 10 min, incubation at room temperature for 60 min with peroxidase-conjugated swine anti-(rabbit IgG) antibodies diluted 1 :20 (Dakopatts, code no. P 217) in buffer B followed at room temperature for

737

Fig. 1. Immunoblotting analysis of conditioned culture medium from cells with anti-tetranectin IgG. Serum-free conditioned culture medium from WI-38 cells was concentrated 380-fold by ultrafiltration. An amount corresponding to 38 ng tetranectin-immunoreactive material, as measured by ELISA (B) and 20 ng purified plasma tetranectin (C) were applied to 10% PAGE. The proteins were transferred on to nitrocellulose paper and incubated with affinity-purified rabbit polyclonal anti-tetranectin IgG followed by phosphatase-conjugated swine anti-(rabbit IgG) antiserum. The positions of prestained M , markers are indicated in (A)

60 min; the washing with buffer A was repeated, followed by an incubation for 60 min at room temperature with peroxidase -anti-peroxidase-antibody complex diluted 1:40 (Dakopatts, code no. Z 113) in buffer B. After another wash with buffer A, twice for 10 min, the preparations were developed in 3-amino-9-ethylcarbazole for 10 min and counterstained with Mayer's hematoxylin for 5 min, both at room temperature. Glycergel (Dakopatts) was used as mounting medium. To ascertain that staining was specific, some cells were incubated without primary specific antibody, and with some cells the primary antibody was replaced with antitetranectin IgG adsorbed to purified tetranectin. Antibodies against prostate-specific antigen (Dakopatts, code no. A 562) and von Willebrand factor (factor-VIII-related antigen; Dakopatts code no. A 082), were used as negative controls. As a positive control for staining of extracellular matrix, a rabbit polyclonal antibody against fibronectin (Dakopatts code no. A 245; IgG concentration, 10.0 g/l) was used. Solubilization of extracellular the matrix Conditioned medium was harvested from dishes seeded with WI-38 and removal of the cells was performed as described above. Extracellular matrix was removed using a rubber policeman and solubilized in the presence of 11% SDS. Solubilized extracellular matrix was analysed by immunoblotting analysis. RESULTS Secretion of tetranectin-immunoreactive material by WI-38 cells In order to screen for production of tetranectin-immunoreactive material in a number of cell lines, their conditioned media were analysed by ELISA. This analysis revealed that WI-38 cells secreted tetranectin-immunoreactive material into the medium at a concentration of 0.7- 1.2 pg . 1-' . cells

Fig. 2. Affinity chromatography of WI-38 conditioned culture medium on plasminogen kringle-4 - Sepharose 4B,followed by SDSjPAGE and immunoblotting analysis for tetranectin-immunoreactive material. (A) Conditioned medium; (B) flow-through and (C) adsorbed material which was released with 1 mM 6-aminohexanoic acid. For further details, see Materials and Methods

Fig. 3. Affinity chromatography of WI-38conditionedmediur I on rabbit anti- (human tetranectin)-IgG - Sepharose 4Bfollowed by SL'SIPAGE and immunoblotting analysis for tetranectin-immunoreactivematerial. (B) Proteins passing through the column; (C) absorbed tetranectinimmunoreactive material which was released from the column with 3 M KSCN. (A) Prestained molecular markers. For further details, see Materials and Methods

.48 h- l. A number of other cell lines, including HT-1080, He1 299, HL 60, HeLa, Hep 2 and HFS 6 cells displayed very little or no tetranectin-immunoreactive material (unpublished results). In order to obtain further evidence for the relationship of the material secreted by WI-38 cells to tetranectin, concentrated conditioned culture medium was subjected to immunoblotting analysis with affinity-purified anti-tetranectin IgG. Two immunoreactive bands around MI 17000, and an additional band with M I 100000, were revealed (Fig. 1B). The M , 17000 bands had a position identical to the positions of the two bands of tetranectin purified from plasma (Fig. 1C). The relative amounts of the two M , 17000 bands varied

738

Fig. 4. Crossed imniLinoelectrophoresis ut p H 8.6 of purified plusma tetranectin and WI-38 tetrunectin-immunareuctive rnatrriul purified b,v immunouffinity chromatography. Samples applied in the slit for the first-dimension electrophoresis were 1 pg purified human plasma tetranectin (A) and 140 ng purified telranectin-immunoreactive material from WI-38 conditioned medium (B - D). First-dimension electrophoresis (anode to the right) was run in the absence ( A - B) or presence of 2 mM CaClz (C) in the electrophoresis buffer, and heparin. 2667 Ulml in the gel (D). Second-dimension elcctrophoresis (anode at top) was run in agarose containing IgC against human tetranectin, at an IgC concentration of 25 mg/l (A) and 12.5 mg/l (B - D)

slightly from experiment to experiment. No staining was seen when the affinity-purified anti-tetranectin IgG was replaced by non-immune rabbit IgG (not shown). SDSjPAGE and Coomassie blue staining of conditioned culture medium supernatant showed a number of protein bands, most with high M , (not shown). Application of the conditioned medium to kringle 4 immobilized on Sepharose 4B showed that most of the MI 17000 protein bound to the column and could be released with 6-aminohexanoic acid (Fig. 2B), whereas the M , 100000 protein passed through the column together with small amounts of the M , 17000 protein (Fig. 2A). Reapplication of the flow-through to the column resulted in binding of a minor part of the low-M, protein; however, a small fraction still passed the column. Tetranectin-immunoreactive material purified from cell medium by anti-tetrdnectin-IgG - Sepharose 4B contained the M , 17000 protein, while the M , 100000 band was absent (Fig. 3). Western blotting of the proteins which passed through the antibody column did not reveal any tetranectinimmunoreactive material, suggesting that the MI 100000 protein was either degraded during purification or not released from the column. The loss of the M , I00000 protein was not prevented by chromatography in the presence of 10 pM aprotinin [17].

Crossed immunoelectrophoresis of the preparation in the presence of anti-tetranectin IgG showed that the electrophoretic mobility of the tetranectin-immunoreactivematerial was influenced by the presence of CaCl, as well as EDTA in the buffer of the first dimension electrophoresis (Fig. 4 A - C), as previously described for plasma tetranectin [l]. Heparin incorporated into the agarose gel for the first-dimension electrophoresis also influenced the electrophoretic mobility (Fig. 4D), as previously described for plasma tetranectin [l]. On the basis of these results, we conclude that a molecule, which is indistinguishable from plasma tetranectin by several criteria, is released by WI-38 cells into the medium. Below we will refer to this molecule as tetranectin-related protein. In addition, the conditioned medium contains a band with higher M , that also reacts with the anti-tetranectin IgG preparation. The nature of this band is unknown. Tetranectin-relatedprotein in the extracellular matrix ,from WI-38 cells

Solubilized matrix derived from WI-38 cells, defined as the material remaining bound to the substratum after removal of the cells, also contained the M , 17000 protein, as identified by immunoblotting (Fig. 5A). A 5-h incubation of the dishes with NaC1/Pi, after removal of the cells, lead to release of part

739

Fig. 5. Immunoblotting analysis of solubilized extracellular matrixfrom WI-38 cultures with anti-tetranectin IgG. The cells were removed from cultures grown in 15-cm-diameter dishes, as described in Materials and Methods. The remaining extracellular matrix was incubated for 5 h with 20 ml NaCl/P, in each plate, after which time the NaCl/P, was removed and concentrated 100-fold, then the extracellular matrix was solubilised in I .O ml 11% SDS. Solubilised extracellular matrix (A) and concentrated NaCl/P, from the 5 h incubation (B) were subjected to SDSjPAGE and immunoblotting analysis with affinitypurified rabbit polyclonal anti-tetranectin IgG

of the M I 17000 protein, as shown by immunoblot analysis of concentrated buffer (Fig. 5B).

Immunocytochemical staining of tetranectin-related protein in WI-38 cells and extracellular matrix Immunocytochemical analysis of WI-38 fibroblasts for tetranectin-immunoreactive material showed a fine granular, cytoplasmatic staining of all cells (Fig. 6A). The staining varied from cell to cell within the same preparation. The preparation of extracellular matrix, which emerged after removal of the cells, was distinctly stained, forming fine strands (Fig. 7). As a control, cells incubated with anti-(prostatespecific antigen) IgG (Fig. 6C) and cells incubated with adsorbed anti-tetranectin IgG (not shown), as well as cells incubated with anti-(von Willebrand factor) IgG (not shown), did not show any staining. The same result was obtained when the primary specific antibody was excluded from the staining procedure. Analysis of the cells for fibronectin (positive control) showed a weak cytoplasmic staining of the cells, and an intense staining of extracellular matrix strands (Fig. 6 B). The demonstration of the MI 17000 bands, and no other immunoreactive bands, by immunoblotting analysis of the solubilized extracellular matrix, serves as a control for the specificity of the staining. DISCUSSION In this study, we demonstrate that a tetranectin-related protein is produced by the human embryonic lung fibroblast cell line WI-38, deposited in extracellular matrix on the growth

substratum and released into the culture medium. The study was performed with cells which were kept under serum-free conditions after having reached confluence. Under such conditions, the cells do not multiply, therefore we avoided the presence of fetal calf serum which might disturb quantification by ELISA and complicate protein purification. Identification of tetranectin-related protein in these cells was based on the observation of a protein with a number of properties in common with tetranectin purified from plasma. The proteins not only shared immunoreactivity and M,, but electrophoretic mobility was also, in both cases, affected by CaCl,, EDTA and heparin [l].Furthermore, the plasminogen-kringle-4-binding ability was present in the protein from WI-38 cells as well as in plasma tetranectin. We cannot, however, exclude the possibility that small differences exist between fibroblast tetranectin-related protein and plasma tetranectin, with respect to amino acid sequence and/or post-translational modification. In addition to the M I 17000 tetranectin-related protein, a M , 100000 protein was observed by immunoblotting analysis of conditioned culture fluid with an affinity-purified antibody raised against plasma tetranectin. The nature of this highM , protein remains unknown. It did not bind to kringle-4Sepharose. One may therefore speculate that it is a precursor of tetranectin with a hidden kringle-4-binding site. The relative amounts of the MI 17000 and the M I 100000 proteins varied from one conditioned medium to the other, and the M I 100000 protein often disappeared during purification procedures. This apparent loss of the high-MI protein is compatible with it being degraded to the M , 17000 tetranectin-related protein during handling of the media. Alternatively, it may be an irrelevant protein which has contaminated the antigen preparation used for antibody production. Further studies are needed to distinguish between these two possibilities. Tetranectin-related protein was found not only in the conditioned cell medium, but was also associated with the extracellular matrix. Extracellular matrix is defined as the material which remains bound to the growth substratum after removal of the cells, and is expected to contain both extracellular matrix components produced by the cells, material from the calf serum in the culture medium, which adhered to the glass or plastic surfaces, and cellular contact site components (for references, see [18]). Application of fetal calf serum to an anti-tetranectin-IgG-linked Sepharose, followed by elution of adsorbed proteins, did not result in any tetranectin-related material by analysis in immunoblotting with rabbit antitetranectin IgG (unpublished results). This shows that putative bovine tetranectin did not react with a rabbit antibody against human tetranectin. The deposition of the tetranectin-related protein on the extracellular matrix might occur prior to its accumulation in the conditioned medium. It is possible that tetranectin-related protein, after secretion from the cells, is deposited on the substratum and later released into the conditioned medium, as reported for plasminogen-activator-inhibitor 1 and thrombospondin [18-201. The idea is supported by the fact that tetranectin-related protein is released from the extracellular matrix into the buffer, which was been added after removal of the cells. Tetranectin has been found to bind to sulphated polysaccharides [21], common components of proteoglycans of connective tissue. As fibroblasts synthesize proteoglycans [22], it is possible that proteoglycans deposited into the extracellular matrix under the WI-38 cells are responsible for binding the tetranectin-related protein at that location.

740

Fig. 6. hiniunoc.~toc.hemic.(~I demonstration of tetrunectiti-relatedprotein in WI-38 cells. WI-38 cells were seeded on slide flasks and stained for tetranectin (A), fibronectin (B) and prostate-specific antigen (C) by the indirect immunoperoxidase methods and counterstaining with Mayer’s hematoxylin (magnification, x 500). In (A) and (B), the reaction products appears as randomly distributed fine granules within the cytoplasm and as fine strands extracellular (arrows). No staining was seen when anti-(prostate-specific antigen) antibody was applied (C)

741

Fig. I. Indirect immunoperoxidase staining for tetranectin-related protein of strands of extracellular matrix which remained on the slide jlasks after removal of cells. A distinct, tetranectin-positive reaction is apparent within the strands (magnification, x 350)

The biological function of tetranectin remains to be determined. Its presence in a variety of secretory cells [15, 231 and its release from neutrophil granulocytes upon stimulation [24] suggest that one of its functions is related to secretion. The present findings suggest other functions of tetranectin and/ or tetranectin-related proteins. The deposition of tetranectinrelated protein by WI-38 cells into the extracellular matrix on the growth substratum in the cultures suggests that it may also be present in the extracellular matrix in vivo. Recently, tetranectin immunoreactivity was demonstrated in many breast carcinomas in the extracellular matrix surrounding the tumor cells, which often had reduced content of tetranectin [25].Extracellular tetranectin in breast carcinomas was characteristically present when the tumor displayed a proliferative connective tissue response, the so-called desmoplastic reaction [26]. An increased number of large fibroblasts (termed myofibroblasts) is found in this desmoplastic tissue, which in the light of the present study may account for tetranectin-related protein production. Future studies should be aimed at characterizing the components of the extracellular matrix to which tetranectin-related protein is bound, and the possible function of tetranectin as a plasminogen anchor at that position. Immunohistochemical studies with increased sensitivity may help to establish whether tetranectin is present in extracellular matrix in general. The expert technical assistance of Mrs Kirsten Lund Jakobsen, Mrs Anna Lise Ryberg, Mrs Anne Christensen and Mrs Inge Beck is gratefully acknowledged. The work was supported by the Danish Cancer Society and the Danish Medical Research Council.

REFERENCES 1. Clemmensen, I., Petersen, L. C. & Kluft, C. (1986) Eur. J . Biorhem. 156, 327-333. 2. Fuhlendorff, J., Clemmensen, I. & Magnusson, S. (1987) Biochemistry 26, 6751 -6164. 3. Spiess, M., Schwartz, A. L. & Lodish, H. F. (1985) J . Biol. Chem. 260, 1979-1982.

4. Spiess, M. & Lodish, H. F. (1985) Proc. Natl Acad. Sci. USA 82, 6465 - 6469. 5. Drickmar, K., Manon, J. F., Bins, G. & Leung, J. 0. (1984) J . Bid. Chem. 259, 710 - 178. 6. Drickmar, K. (1983) J . Biol. Chem. 256, 5827-5839. I. Doege, K., Fernandez, P., Hassel, J. R., Sasaki, M. & Yamada, Y. (1986) J . Biol. Chem. 261, 8108-8111. 8. Sai, S., Tanaka, T., Kosher, R. A. & Tanzer, M. L. (1986) Proc. Natl Acad. Sci. U S A 83, 5081 -5085. 9. Drickmar, K., Dordal, M. S. & Reynolds, L. (1986) J . Biol. Chem. 261,6818-6887. 10. Liidin, C., Hofstetter. H., Sarfati, M., Levy, C. A., Suter, U., Alaimo, D., Kilchherr, E., Frost, H. & Delpresse, G. (1987) EMBO J. 6, 109-114. 11. Jensen, B. A,, McNair, P., Hyldstrup, L. & Clemmensen, I. (1987) J . Lab. Clin.Med. 110, 612-617. 12. Jensen, B. A. & Clemmensen, I. (1988) Cancer 62, 869-872. 13. Lund, L., Georg, B., Nielsen, L. S., Mayer, M., Dan@, K. & Andreasen, P. A. (1988) Mol. Cell. Endocrinol. 60, 43-53. 14. Burton, K. (1956) Biochem. J . 62, 315-321. 16. Christensen, L., Johansen, N., Jensen, B. & Clemmensen, I. (1987) Histochemistry 87, 195 - 199. 16. Bjerrum, O., Ramlau, J., Clemmensen, I., Hansen, T. B. & Ingild, A. (1975) Scand. J . fmmunol. 2 (Suppl.) 81 -85. 11. Thorsen, S. & Mullertz, S. (1974) Scand. J . Clin. Lab. Invest. 34, 167 - 176. 18. Pollanen, J., Saksela, O., Salonen, E.-M., Andreasen, P. A , , Nielsen, L. S., D a m , K. & Vaheri, A. (1987) J . Cell Biol. 104, 1085 - 1096. 19. Knudsen, B. S., Harpel, P. C. & Nachmann, R. L. (1987) J . Clin. Invest. 80, 1082- 1089. 20. Jaffe, E. A,, Ruggiero, J. T., Leung, I. L. K., Doyle, M. J., McKeown-Longo, P. J. & Mosher, D. F. (1983) Proc. Natl Acad. Sci. USA 80,998 - 1002. 21. Clemmensen, I. (1989) Scand. J . Clin. Lab. Invest. 49, 719-125. 22. Iozzo, R. V., Bolender, R. P. & Wight, T. N. (1982) Lab. Invest. 47,124-138. 23. Christensen, L. & Clemmensen, I. (1989) Histochemistry 92,2935. 24. BorregBrd, N., Christensen, L., Bjerrum, 0. W., Birgens, H. & Clemmensen, I. (1990) J . Clzn. Invest. 85,408-416. 25. Christensen, L. & Clemmensen, I. (1991) Histochemistry (in press). 26. Lagace, R., Grimand, J.-A,, Schwurch, W. & Seemayer, T. A . (1985) Virchows Arch. (Pathol. Anat.) 408,49- 59.