Exp Brain Res (1992) 88:213-218

E erimental BrainResearch © Springer-Vertagt992

Synthesis and secretion of 2-macroglobulin by human glioma established cell lines R. Businaro 1, C. Fabrizi 2, L. Fumagalli 1, and G.M. Lauro 2 1 Dipartimento di Scienze Cardiovascolari e Respiratorie, and 2 Dipartimento di Biologia Cellulare e dello Sviluppo, Universitfi "La Sapienza", Via degli Apuli n° 1, 1-00185 Roma, Italy Received March 4, 1991 / Accepted August 15, 1991

Summary. Human a2-macroglobulin (a2M) is a high molecular weight plasma proteinase inhibitor exhibiting a broad specificity; in fact it is capable of binding endopeptidases from all known classes of proteases (Barret 1981). Two human glioma cell lines, namely an astrocytoma and a glioblastoma, were found to synthesize and secrete in the culture medium a protein which resembles the serum a2M for immunological, biochemical and biological features. Using polyclonal antibodies to serum (~2M, an a2M-like factor could be detected in the cytoplasm and in the culture medium of the tumor cells. Furthermore this factor accumulated in cytoplasmic granules if cells were incubated with monensin and its production was dramatically reduced following a treatment with cycloheximide. This protein behaved like the serum o~2Min immunoblotting analysis and exhibited the same antiproteolytic activity. Its role in human brain is unknown at present. Since interactions of proteinases and proteinase-inhibitors appear to influence the hosttumor immune response and to play a crucial role during the migration of metastasizing tumor cells, a2M expression observed in these glioma cells could be involved in tumor cell proliferation and invasion. Key words: Gtial cultures - Gliomas Tumor cells Proteinase inhibitors - a2-macroglobulin - Human

ment, since many aspects of mitogenesis, migration and cell movement are controlled by proteases. It has been also shown that oncogene activation, which induces cell proliferation and malignant transformation, leads to the proteinase gene expression (Tryggvason et al. 1987). Furthermore interactions of proteinase and proteinaseinhibitors appear to play a crucial role during brain development, in cell migration (Guenther et al. 1985; Lindner et al. 1986) and neurite outgrowth (Pittman 1985; Hawkins and Seeds 1986; Monard 1988). In this connection, recent experiments by Mori et al. showed that ~2M derived from newborn rat astroglial cells induce neurite elongation in cultured neurons from rat central nervous system (Mori et al. 1990). In the present study we decided to evaluate the presence of proteinase inhibitors in two human glioma established cell lines, namely an astrocytoma and a glioblastoma. Occurrence of some of these inhibitors, such as c~2-macroglobulin (~2M), in human brain has been already reported (Dziegelewska et al. 1986) and its presence has been demonstrated immunohistochemically in human cerebral gliomas (Seitz and Wechster 1987). Its synthesis was observed in vitro in astrocyte primary cultures from newborn rats (Gebicke-Haerter et al. 1987). Moreover, several human melanoma and sarcoma cell lines synthesize and secrete ~zM (Saksela et al. 1984; Bizik et al. 1986; Grofova et al. 1988; Lizonova et at. 1990) and recently a glioblastoma multiforme was found to produce ~zM in an organ culture system (Keohane et al. t990).

Introduction Increased levels of proteinase inhibitors were recently found in neoplastic tissues and sera of tumor-bearing patients (Harris et al. 1974; Cooper et al. 1976; Matsuda et al. 1983; Okumichi et al. 1984; Cheung and Lau 1986; Chawla et al. 1987; Sawaya et al. 1987). It is well known that the metastatic capacity may be influenced by the proteolytic activity present in the extracellular environOffprint requests to: G.M. Lauro

Material and methods Cell lines

Two human glioma cell lines, obtained from explants of a III WHO gemistocytic astrocytoma (T67; 27th-30th passage in culture) and from a glioblastoma (T70; 23rd-26th passage), were grown routinely in monolayers within 25-cm2 flasks (Nunc, Denmark) containing HAM-F10 (Gibco, Paisley, UK) supplemented with 10% fetal calf

214 serum (FCS; Gibco, Paisley, UK) and gentamycin (40 gg/mt; Hazleton, USA). Cells were cultured in a fully humidified atmosphere of 5 % CO2 in air and passaged with 0.05 % trypsin and 0.02 % EDTA (Flow, Irvine, UK) at a split ratio of 1 : 2. For all experiments cells were washed twice with phosphate buffered saline (PBS) and then placed in serum free medium (SFM) for at least three days. Both lines (T67, T70) were characterized by means of polyclonal and monoclonal antibodies directed against glial fibrillar acidic protein (GFAP), S-100 protein, fibronectin, factor VIII and vimentin (Dakopatts, Denmark) as previously described (Lauro et al. 1986; Cusimano et al. 1990).

Proteinase~inhibitors detection by immunofluorescence T67 and T70 cells were grown on coverslips for five days in SFM, washed extensively in PBS, fixed 10 rain in acetone at - 2 0 ° C, washed again and then incubated overnight at 4° C in a moist chamber with the various rabbit polyclonal antibodies directed against: ctzM, 1: 200; inter-c~-trypsin inhibitor, 1 : 200; a 1-antitrypsin, 1: 200; ctx-antichymotrypsin, 1:200 (Dakopatts, Denmark). Then they were treated for 1 h at room temperature with a mouse monoclonal anti-GFAP antibody diluted 1: 10 (Boehringer Mannheim, FRG). After removing the specific antiserum and several washings with PBS, the cells were treated, first with a rhodamine conjugated anti-rabbit IgG F(ab')2 antibody (30 min; diluted 1:50), and then with a ftuorescein conjugated anti-mouse IgG F(ab')2 antibody (30 min; diluted 1:50), both obtained from goat (Cappel, Cochranville, PA, USA). The cells were finally washed with PBS, mounted on glycerol-PBS and observed with a Leitz Dialux 20 fluorescence microscope under the appropriate wavelengths.

resolving gels and 3 % stacking gels. Proteins were visualized using a silver stain procedure (Schoenle et al. 1984) or transferred to nitrocellulose membrane (Bio-Rad, CA, USA) for immunoblotting (Towbin et al. 1979). The membrane was incubated first with rabbit antibodies anti-azM (Dakopatts, Denmark) diluted 1:100, then with donkey biotinylated anti-rabbit Ig (Amersham, Bucks, UK) and finally with streptavidin-biotinylated horseradish peroxidase complex (Amersham, Bucks, UK). The bands were visualized using 4-chloro-l-naphthol (Sigma, St. Louis, Mo, USA) as substrate.

ELISA Enzyme-linked immunosorbent assay (ELISA) was performed essentially as described by Engvall et aL (1971). Briefly, 96-well ELISA plates (Nunc, Denmark) were incubated with: a) a2M purified from human serum, used as a positive control; b) medium obtained from cultures maintained in SFM for 72 h, as previously described. After incubation overnight at 4° C, plates were coated with 1% BSA in PBS, treated, first with polyclonal rabbit antibodies directed against ctzM, ch-antitrypsin or ctl-antichymotrypsin, then with donkey biotinylated anti-rabbit Ig (Amersham, Bucks, UK) and finally with streptavidin-biotinylated horseradish peroxidase complex (Amersham, Bucks, UK). A solution containing 0.04% o-phenylenediamine(Sigma, St.I,ouis, Mo, USA), 0.1 M citric acid, 0.2 M NazHPO4, and 0.012% H202 was used as substrate. The absorbance was read at 492 nm on an automated ELO 310 ELISA reader (Biotech, Cambridge, Ma). All plates were blanked against antigen coated wells which were incubateA with diluent only or rabbit normal IgG. The unknown concentration of azM in the samples was determined using a standard curve obtained with scalar dilutions of sermn ctzM as described by Kramer and Justus (1988).

Cycloheximide treatment Monensin treatment T67 and T70 cells, grown on coverslips, were maintained for five days in SFM, then replaced for 2 h with fresh SFM containing 10 gM monensin (Sigma, St.Louis, MO, USA); finally, they were washed in PBS, fixed and immunohistochemically processed as previously described.

Isolation and purification of c~zM T67 cells were grown to confluency (85 cm 2 flasks; Nunc, Denmark) in F10, 10% FCS and antibiotic. After an incubation for 72 h in SFM, they were washed twice in PBS in order to remove residual serum proteins adsorbed to the cell layers and replaced with fresh SFM (20 ml per flask). After 72 h the culture medium was collected, passed through a 0.22 pm filter to remove cellular debris and stored at - 20° C. Five hundred ml of culture medium were applied to an affinity cromatography column packed with Sepharose-4B by the method of Parikh et al. (1974). The Sepharose was activated with cyanogen bromide and then coupled to a rabbit antibody directed against azM (Dakopatts, Denmark). Bound protein was eluted with a linear gradient of 0 to 4.5 M MgC12. Fractions (2 ml) were collected and measured for opticat absorbance at 280 nm. The protein peaks were pooled and dialyzed against distilled water. The same procedure was used for purification of tIzM from human serum obtained from healthy donors. The concentration of azM was determined based on A2so nm 1% =9.0 (Barret 1981).

SDS-PAGE and Western blotting a2M purified from glioma culture medium or from human serum was run in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) according to Laemmli (1970) using 7.5%

Cells were cultured in SFM for 72 h; then the culture medium was discarded and the cells incubated with cycloheximide (50 gg/ml; Sigma, St. Louis, Mo, USA) diluted in fresh SFM. In the controls, only SFM was added. After 3 or 6 h, medium was aspirated from the cultures, cells briefly trypsinized, diluted in PBS, and immediately counted. Prior to counting, the cells were examined by light microscopy for morphological appearance and viability. As assessed by morphological criteria and dye (Trypan blue) exclusion, attached cells were > 95% viable.

Antbtrypsin activity assay Since both 1:1 and 2:1 proteinase-uzM complexes can be formed, a double molar ratio trypsin-a2M was used to obtain a complete saturation of u2M. Tripsyn-ct/M complex was prepared by mixing 1 tool of uzM with 2 tool of trypsin (Sigma, St. Louis, Mo, USA) at 37° C for 30 rain. All proteins were dissolved in 0.1 M Tris HC1 pH 7.5, 2 mM CaCI2. l'*C-methylated casein (Sigma, St. Louis, Mo, USA) was added as substrate (0.08 ~g in 20 ~tl of 10 mM PBS) and incubated 30 rain at 37° C. The reaction was stopped at 4° C, the solution precipitated with thrycloracetic acid (20% w/v) and then centrifuged at 6000 g for 15 rain. The radioactivity present in the supernatant was determined using a liquid scintillation counter (Packard 360 C). The percentage of inhibition was calculated comparing the values obtained in the presence of trypsin alone with those measured in the presence of the trypsin-azM complex.

Results

Cytological markers Both T67 a n d T70 cell lines were positive for G F A P (which is a specific m a r k e r for astroglial cells), S-100 p r o t e i n a n d v i m e n t i n (which c a n be expressed by astro-

215

Fig. 1A, B. Double-immunofluorescencestaining of T67 cells for GFAP A, and for ~2M B showing their colocalization in the same cells. Bar indicates 15 pm

Fig. 2. Immunocytochemicatdetection of c~zMin T67 cells treated with monensin (10 gM) for 2 h: the positivity observed is granular and exclusivelyperinuclear. Bar indicates 10 lain 1200-

gtial cells); no positivity was detected for fibronectin (a typical marker for fibroblasts) and factor VIII (a specific marker of endothelial cells).

[XX] cycloheximide

~.

Gtioma derived proteinase inhibitors Cells treated with anti-a2M antibodies showed in indirect immunofluorescence a strong positivity occurring in about 80% of the cells. Figure 1 shows the colocalization of GFAP (Fig. la) and ~2M (Fig. lb) in T67 cells; in particular, a2M related immunoreactivity was peculiarly granular and occurred in the perinuclear region and some elongations. When monensin was used the reaction was stronger and exclusively perinuclear (Fig. 2). Very similar findings, but of weaker intensity, were obtained using the T70 cell line. In both lines, no positivity was observed following incubation with antibodies directed against: inter-a-trypsin inhibitor, al-antitrypsin and al-antichymotrypsin. Furthermore c~zM was detected in the culture medium by ELISA and its secretion could be prevented using an inhibitor of protein synthesis, cycloheximide (Fig. 3 a, b). Figure 4 depicts the a2M secretion during four days of culture: most of it occurred during the first 24 h, while only a little increment was observed in the following days (from 2nd to 4th). Similar to immunocytochemical data, T67 (astrocytoma derived cell line) seemed to produce larger amounts of this protein compared with T70 (glioblastoma derived cell line), maintained in the same culture conditions (Fig. 4). In particular, a2M detected in T67 cell line culture medium was about 5-fold compared with T70 cell line. The culture supernatants assayed for the presence of other plasma proteinase-inhibitors were found to be negative for eq-antitrypsin and ch-antichymotrypsin. The a2M-like molecule purified from T67 culture medium by affinity cromatography was analyzed by SDS-polyacrylamide gel electrophoresis and compared with a2M obtained from human serum. In both cases, only one prominent band of apparent molecular weight

1

1000-

A T67 cells

800600-

&

400200O5 1200

I

1000E

6

[XZ] cycloheximide m control

[3 T70 cells

800-

c

[

60040020000

3 Time (hours)

Fig. 3A, B. Cycloheximide blocking effect on c~2Mrelease by T67 (A) and TT0 (B) cells during serum-free maintenance. Cycloheximide (50 gg/ml) was added to cellsfor a period of 3 or 6 hours. Each point refers to 5 x 105 cells and represents the mean value:ks.& from two flasks

of 180 Kd could be observed. For serum a2M some minor bands, probably representing degradation products (Barret 1981), were also present. The immunoblotting analysis using anti-azM antibodies confirmed these results (Fig. 5). Successively, the anti-trypsin activity of the tumor derived a2M-like molecule was determined: it decreased the proteolytic effect of trypsin by 85% (Fig. 6). No

216 60005000. E

% Inhibition of t r y p s i n I~

T67 cells

1

T70

100 -

-k -k

90-

4000" 80 3000.

70-

I

t3

2000.

60-

,< 1000" O' 0

50-

• ,,I

Ill, 48

24

40 72

96

Time (hours)

Fig. 4. Release of a2M by T67 and T70 cells observed during 4 days of culture. Each point refers to 5 x 105 cells and represents the mean value :k s.d. from two flasks

30.

20 10 o-

A

B

200 K d 116 K d 97 K d 66 K d 43 K d -

Fig. 5. Western blot analysis of ct2M. Lane A : c~2Mpurified from human serum; lane B: % M purified from T67 culture medium. The position of the molecular weight markers (Bio-Rad) is shown in the left margin: myosin, 200 Kd; [3-galactosidase, 116 Kd; phosphorylase b, 97 Kd; bovine serum albumin, 66 Kd; ovalbumin, 43 Kd

significative difference was observed in this assay between ~2M derived from glioma culture medium and that obtained from serum.

Discussion Our results indicate that two human glioma cell lines (T67, TT0) synthesize and release in the culture medium an a2M-like molecule which has been characterized by biochemical, immunochemical and biological assays. By indirect immunofluorescence a2M was observed both in untreated cells and following incubation with monensin, a protein secretion inhibitor (Tartakoff and Vassalli 1978). When monensin was used, the reaction was stronger and localized at the level of cytoplasmic granules. These granules probably correspond to Golgi cisternae, organelles where secretory products accu-

1

Serum a2M

[~

Glioma-derived a2M i

Fig. 6. Inhibitory effect on trypsin proteolytic activity of c~2M purified from human serum or from glioma culture medium. Statistical comparison by the Student's t-test showed no significant differences between serum and glioma derived ~zM (*)

mulate, suggesting de-novo synthesis of Gt2M. This point is confirmed by the lack of a2M in the culture medium of cells treated with cycloheximide, a well known inhibitor of protein synthesis. On the other hand, since in all our experiments fetal calf serum was removed by extensive washing and by incubating the cultures for at least 72 h in serum free medium, it seems unlikely that ~2M detected here comes from serum. Both lines (T67, T70) assayed by immunofluorescence for the presence of a2M showed similar findings but of different intensity. The fact that not all gtioma cells were stained by anti-a2M antibodies and that the positive cells did not give a signal of similar intensity, can possibly be explained by an absence of synchronization in our cultures. In fact, cells entering different phases of the cellcycle undergo a modulation of protein expression (Dulbecco et al. 1983). According to immunofluorescence results, very different amounts of the protein were detected by ELISA in the culture medium of the two lines. In particular, T67 cell line produced about 5-fold more a2M than the T70 cell line, maintained in the same culture conditions. Perhaps, it could be linked to the different grade of malignancy of the tumors from which these lines derived. Recently the synthesis of a2M by a human glioblastoma multiforme tissue, using an organ culture system was reported (Keohane et al. 1990). The rate of secretion of this and other proteinase-inhibitors was found to increase significantly with time, correlating with immunohistochemical evidence of differentiation induced by this method of culturing (Rubinstein and Herman 1975). Our biochemical data show that the a2M-like factor purified from T67 cell supernatants has the same molec-

217 ular weight as ~2M obtained f r o m h u m a n serum (180 Kd). On the contrary, some Authors observed that ~zM derived from m e l a n o m a and sarcoma cell lines exhibited a molecular weight of 140 Kd, p r o b a b l y related to differences at the level o f glycosilation ( G r o f o v a et al. 1988). Both serum- and tumor-derived ~2M are bound by anti-a2M antibodies, as shown by E L I S A and western blotting; moreover, the c%M like factor isolated from gtioma culture m e d i u m is functionally active; in fact it exhibits a strong inhibitory function against trypsin proteolytic activity, paralleling the effect of serum a2M. However, further molecular characterization of the gtioma derived protein is required to define its h o m o l o g y to serum % M . The role and significance of this t u m o r derived proteinase inhibitor are not known at present. Malignancy, and in particular invasiveness, have often been linked to increased proteolytic activity. Considerable evidence has amassed suggesting that protease production by t u m o r cells m a y be required for invasiveness (Tryggvason et al. 1987). Conversely, evidence suggests that some proteinase inhibitors, such as a2M m a y modulate the host-tumor immune responses and have immunosoppressive effects on natural killing and antibody dependent cytotoxicity (Dickinson et al. 1985), polyclonal B-cell activation (Chang et al. 1983), and the mixed lymphocyte response ( H u b b a r d et al. 1981). In this connection, a bad prognosis has been suggested for patients with ~2M containing tumors (Tahara et al. 1984; M a t o s k a et al. 1988; K a t a o k a et al. 1989). Recently, serum % M was shown to bind to a wide range o f cytokines, including nerve growth factor, suggesting a potentially important mechanism for the regulation of their activity (James 1990; R o n n e et al. 1979). With reference to this, a m o d u l a t o r y effect of a2M on the nervous system cannot be excluded. Moreover, since interactions of proteinases and proteinase-inhibitors appear to play a crucial role in cell migration (Guenther et al. 1985; Lindner et al. 1986) and neurite outgrowth (Pittman t985; Hawkins and Seeds 1986; M o n a r d 1988), % M synthesis by h u m a n astrocytes could be involved in brain developmental events. Our preliminary results indicate that h u m a n fetal astrocytes exhibit intracytoplasmic positivity for a2M. This finding suggest that a2M could have a physiological role during embryo development. In conclusion, these glioma cell lines could be used as a model to try to elucidate the role of a2M, if any, in the growth and differentiation of the nervous cells.

References Barret AJ (1981) %-macroglobulin. Methods Enzymol 80: 737765 Bizik J, Vaheri A, Saksela O, Kalkkinen N, Meri S, Grofova M (1986) Human tumor cells synthesize and secrete %-macroglobulin in vitro. Int J Cancer 37:81-88 Chang JL, Ganea D, Dray S, Teodorescu M (1983) Blast transformation of B celts induced by an e~-macroglobulin associated lymphokine produced in crowed lymphoid cell cultures. J Immunol 130: 267-273

Chawla RK, Lawson DH, Sarma PR, Nixson DW, Travis J (1987) Serum %-proteinase inhibitor in advanced cancer: mass variants and functionally inert forms. Cancer Res 47:1179-1184 Cheung A, Lau HKF (1986) Isolation and partial characterization of a proteinase inhibitor from human colorectal adenocarcinoma. Biochim Biophys Acta 882:200-209 Cooper EH, Turner R, Geeike A (1976) Alphaglobulins in the surveillance of colorectal cancer. Biomedicine 24: 171-178 Cusimano G, Palladini G, Lauro GM (1990) MHC Class II expression by human glioma cells after in vitro incubation with soluble antigens. Acta Neurol Scand 81:215-222 Dickinson AM, Shenton BK, Alomran AH, Donnelly PK, Proctor SJ (1985) Inhibition of natural killing and antibody dependent cell-mediated cytotoxicity by the plasma protease inhibitor %macroglobulin (et2M) and ot2M protease complexes. Clin Immunol Immunopathol 36: 259-265 Dulbecco R, Allen R, Okada S, Bowman M (1983) Functional changes of intermediate filaments in fibroblastic cells revealed by a monoclonal antibody. Proc Natl Acad Sci USA 80:1915-1918 Dziegielewska KM, Saunders NR, Schejter EJ, Zakert H, ZevinZonkin D, Zisling R, Soreq H (1986) Synthesis of plasma proteins in fetal, adult and neoplastic human brain tissue. Dev Biol 115:93-104 Engvall E, Perlmann P (1971) Enzyme linked immunosorbent assay (ELISA). Quantitative assay of immunoglobulin G. Immunochemistry 8: 871-874 Gebicke-Haerter PJ, Bauer J, Brenner A, Gerok W (1987) %-macroglobulin synthesis in an astrocyte subpopulation. J Neurochem 49 : 1139--1145 Grofova M, Larsson E, Bengtsson A, Bizik J, Westermark B, Ponten J (1988) Localization of a2-macroglobutin in human primary sarcomas and synthesis in established cell lines. In vitro Cell Dev Biol 24:369-372 Guenther J, Nick HP, Monard D (1985) A gila-derived neuritepromoting factor with protease inhibitory activity. EMBO J 4:t963-1966 Harris CC, Primack A, Cohen MH (1974) Elevated %-antitrypsin serum levels in lung cancer patients. Cancer 34:280-281 Hawkins RL, Seeds NW (1986) Protease inhibition can influence the direction of neurite outgrowth by neonatal mouse sensory ganglia. J Neurosci (Abstr.) 6:1117 Hubbard WJ, Hess AD, Hsia S, Amos DB (1981) The effects of electrophoretically "slow" and "fast" %-macroglobulin on mixed lymphocyte cultures. J Immunol 126:292-299 James K (1990) Interactions beetween cytokines and %-macroglobulin. Immunology Today 11 : 163-166 Kataoka H, Nabeshima K, Komada N, Koono M (1989) New human colorectal carcinoma cell lines that secrete proteinase inhibitors in vitro. Virchows Arch [B] 57:157-165 Keohane ME, Scott WH, Scott RV, Gonias SL (1990) Secretion of %-macroglobulin, %-antiplasmin, and plasminogen activator inhibitor-1 by gliobtastoma multiforme in primary organ culture. J Neurosurg 73:234~241 Kramer MD, Justus C (1988) The anti-proteolitic compound %macroglobulin in human skin. Arch Dermatol Res 280:93-96 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685 Lauro GM, Di Lorenzo M, Grossi M, Maleci A, Guidetti B (1986) Prostaglandin E2 as an immunomodulating factor released in vitro by human glioma cells. Acta Neuropathol (Berl) 69 : 278-282 Lindner J, Guenther J, Nick HP, Zinser G, Antonicek H, Schachner M, Monard D (1986) Modulation of granule cell migration by a glia-derived protein. Proc Natl Acad Sci USA 83:4568-4571 Lizonova A, Bizik J, Grofova M, Vaheri A (1990) Coexpression of tumor-associated %-macroglobulin and growth factors in human melanoma cell lines. J Cell Biochem 43:31 ~325 Matoska J, Wahlstr6m T, Vaheri A, Bizik J, Grofova M (1988) Tumor-associated %-macroglobulin in human melanomas. Int J Cancer 41 : 359-363

218 Matsuda K, Ogawa M, Murata A, Kitahara T, Kosaki G (1983) Elevation of serum immunoreactive pancreatic secretory trypsin inhibitor contents in various malignant diseases. Res Commun Chem Pathol Pharmacol 40:301-305 Monard D (1988) Cell-derived proteases and protease inhibitors a s regulators of neurite outgrowth. TINS 11 : 541-544 Mori T, Iijima N, Kitabatake K, Kohsaka S (1990) ct2-macroglobulin is an astroglia-derived neurite-promoting factor for cultured neurons from rat central nervous system. Brain Res 527: 55-61 Okumichi T, Nishiki M, Takasugi S, Toki N, Ezaki H (1984) Isolation of urinary trypsin inhibitor-like inhibitor from human lung cancer tissue. Cancer Res 44: 2011-2015 Parikh I, March S, Cuatrecasas P (1974) Topics in the methodology of substitution reactions with agarose. Methods Enzymol 34: 77-102 Pittman RN (1985) Release of plasminogen activator and a calcium-dependent metalloprotease from cultured symphatetic and sensory neurons. Dev Biol 110:91-101 Ronne H, Anundi H, Rask L, Peterson PA (1979) Nerve growth factor binds to serum ~z2-macroglobulin. Biochem Biophys Res Commun 87:330-336 Rubinstein LJ, Herman MM (1975) Studies on the differentiation of human and experimental gliomas in organ culture systems. Recent Results Cancer Res 51:35-51

Saksela O, Vaheri A, Schleuning WD, Mignatti P, Barlati S (1984) Plasminogen activators, activation inhibitors and a2-macroglobulin produced by cultured normal and malignant human cells. Int J Cancer 33:609-616 Sawaya R, Zuccarello M, Highsmith R (1987) cq-antitrypsin in human brain tumors. J Neurosurg 67:258-262 Schoenle EJ, Adams LD, Sammons DW (1984) Insulin-induced rapid decrease of a major protein in fat cell plasma membranes. J Biol Chem 259 :1211~12116 Seitz R J, Wechsler W (1987) Immunohistochemical demonstration of serum proteins in human cerebral gliomas. Acta Neuropathol 73:145-152 Tahara E, Ito H, Taniyama K, Yokozaki H, Hata J (1984) c~i-antitrypsin, al-antichymotrypsin, and az-macroglobulin in human gastric carcinomas: a retrospective immunohistochemical study. Hum Pathol 15:957-964 Tartakoff A, Vassalli P (1978) Comparative studies of intracellutar transport of secretory proteins. J Cell Biol 102:2244-2253 Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrilamide gels to nitrocellulose. Procedure and some application. Proc Natl Acad Sci USA 76:4350-4354 Tryggvason K, H6yhtyfi M, Salo T (1987) Proteolytic degradation of extracellular matrix in tumor invasion. Biochim Biophys Acta 907:19t-2t7