- Email: [email protected]
The estrogen-regulated 52K-cathepsin-D in breast cancer: From biology to clinical applications
Nucl. Med. Biol. Vo1.14, No.4, pQ.377-304, _Int. J.
0883-2897/87 s3.00+0.00
1987
Radiat. Appl. Instrum. Part B
Pergamon
Journals Ltd
Printed in Great Britain
THE
ESTROGEN-REGULATED FROM
Henri ROCHEFORT, Franpoise GARCIA, Muriel MORISSET,
52K-CATHEPSIN-D
BIOLOGY
TO CLINICAL
IN
BREAST
CANCER
:
APPLICATIONS
CAPONY, Patrick AUGEREAU, Vincent CAVAILLES, Gilles FREISS, Thierry MAUDELONDE and Franqoise
Unite d’Endocrinologie Cellulaire et Moleculaire Laboratoire de Biologie Cellulaire and Faculte Navacelles, 34100 Montpellier,
de I’INSERM de Mbdecine, France
(U
60
1481, Rue
Marcel VIGNON and de
SUMMARY We have studied estrogen-regulated proteins in an attempt to understand the mechanism by which estrogens stimulate cell proliferation and mammary carcinoIn estrogen receptor positive human breast cancer cell lines (MCF7, genesis. ZR75-1) estrogens specifically increase the production into the culture medium Several high affinity monoclonal of a 52,000 daltons (52K) glycoprotein. antibodies to the partially purified secretory 52K protein have allowed to purify to homogeneity this protein and its cellular processed products. The 52K protein has been identified as the secreted precursor of a cathepsin-D like via signals and routed to I ysosomes protease bearing mannose-6-phosphate The protease is mitogen ic in vitro on estrogen mannose-6-phosphate receptor. deprived MCF7 cells and is able to degrade basement membrane and proteoglycans The cellular related proteins, as detected by immunofollowing its activation. immunoassay are more concentrated in proliferative mammary histochemistry and in breast cancer cytosol ducts than in resting ducts and their concentration lymph nodes invasion and disease free appears to be more correlated with Copenhagen) than with the estrogen receptor (RE) survival (with S. Thorpe, The protein is also produced constitutively by RE-negative ccl I lines, level. variants, it becomes inducible by whi le in some ant iestrogen resistant Cloning of its cDNA in Xgtll contrary to the wild type MCF7 cells. tamoxifen, has a I lowed to show that the mRNA is rapidly induced by estrogens and to it to that of the norma I human kidney sequence the protein and compare cathepsin-D. We conclude that the estrogen-induced cathepsin-D like protease may in stimulating the growth have important autocrine and/or paracrine functions independent breast cancers and may be and invasion of hormone-dependent and useful as a tissue marker to predict high risk mastopathies and breast cancer In addition to other estrogen regulated growth factors, the invasiveness. the I ysosoma I proteases has the potential to stimulate precursor of pro1 iferation and invasiveness of breast cancer cells as long as they are secreted in excess rather than being routed to lysosomes.
INTRODUCTION The development of human breast cancer cell lines has greatly facilitated the study of the hormonal control of cancer cells proliferation and have allowed to define several proteins which are specifically regulated by steroids. Among those, the proteins and peptides which are induced by estrogens before the growth stimulatory effect of estrogens are particularly interesting since some of them may act as mediator of estrogen action to stimulate the growth and/or invasion of breast cancer cells (for review see l-4). We will summarize the studies from our laboratories which have concentrated on one of these proteins, named 52K from its molecular we i ght in denaturing polyacrylamide gel electrophoresis and have lead to its identification and possible function in carcinogenesis. We will then review the clinical potential of this mammary protein which can be localised and assayed in mammary tissue of patients. BIOLOGY
A.
Regulation
and
OF
THE
52K
PROTEIN
OF MCF7
CELLS
purification
?3he 52K protein was first detected by labeling protein of MCF7 cells with 1’ Slmethionine and analysis of the secreted protein by SDS-polyacrylamide gel electrophoresis (5,6). While estradiol increased by 2 to 3 fold the total amount of secreted proteins, normalisation for constant TCA precipitable material showed that some of the proteins (mainly a 160K and a 52K species) were more increased by estradio I from concentrations as low as 1 to 10 specifically picomole.
377
H. Rochefort et al.
378
into the culture medium by secreted in sm II amounts 52K prote in cells/hour) and by other RE-positive esi!zgen-treated MC$ cells (5 ng/lOb ZR75-1). It is constitutively breast cancer ceils under estrogen control (T47D, in RE-negative cell lines produced without requiring the influence of estrogen the protein is specifically (MDA-MB231, BT20) (7). In RE-positive cells, regulated by hormones (estrogens and high doses of androgens) that can bind to and activate the estrogen receptor, but not by glucocorticoids, progestins, or androgens at low concentrations (6). The effects of the antiestrogens, tamoxifen and hydroxytamoxifen, suggested that this protein was in some way related to the mitogenic activity of estrogens. In wild type MCF7 ccl Is, the antiestrogens total ly inhibited the synthesis of the protein, whereas they partially stimulated the synthesis of the progesterone receptor. By contrast, in the antiestrogen-resistant variants of MCF7 cells (R27 and RTx6) cloned for their ability to grow in 1 PM tamoxifen, the antiestrogens became able, I ike estrogens, to increase the production of the 52K protein, but remained unable, in wi Id type MCF7 cells to stimulate the production of the estrogenZIgulated 160K secreted prote:n and of pS2 mRNA (8). In these cell lines, the 52K protein was therefore a better candidate for being a mitogen than the pS2 or 160K proteins. A possible mechasnism for the ant iestrogen resistance was that ccl Is had acquired a growth advantage, due to the tamoxifen induced increased production of autocrine growth factors such as the 52K protein. Using concanavalin-A sepharose chromatography, we partially purified the 52K protein from 22 liters of conditioned medium from MCF7 cell cultures. From this fraction, we obtained several mouse monoclonal antibodies to the 52K protein Final ly, using an immunoaffinity column, we purified the 52K protein to (9). apparent homogeneity (l,OOO-fold purification) both in its secreted and cellular forms (10). The purification of 52K protein made it possible to study its activity on cell growth, determine its structure and identity, and finally clone its corresponding cDNA sequences. By immunopurification without detergent, a homogeneous secreted 52K protein was obta i ned, which was shown to stimulate the growth of estrogen-deprived recipient MCF7 cells (11). This stimulation was dose-dependent and occurred at concentrations (1 to 10 nM) similar to those found in the culture medium. However, time-course experiments indicated that both estradiol nd the 52K prote in required the same (18 h) lag before stimulating I’Hjthymidine incorporation. The in vitro mitogenic activity of the purified 52K protein was in agreement with an estrogen-regulated autocrine mechanism. To understand such mechanism, it is generally necessary to identify the cell surface receptor which is triggered by the mitogen. In our case, we first tried to identify the 52K protein. 6.
Structure
and
identification
as
a cathepsin-D
Study of the coand post-translational modifications of the 52K protein helped us to define i% structure and enzymatic activity. After exposure of cultured MCF7 cells to P, the 52K protein is intensely labeled. However, most of this label can be removed by endoglycosidase-H treatment, which deletes the two N-glycosylated chains of the protein. Mannose-6-P signals were then identified on these chains (12). Pu I se-chase exper i ments and Western b I ot analysis showed that the 52K protein is the precursor of a lysosomal enzyme which accumulates in lysosomes as a stable 34K protein (Fig. 1) (13). About 40% of the ccl lul ar 52K precursor is secreted while about 60% is successively processed into a 48K and a 34K + 14K protein. Part of the secreted 52K protein can be taken up and processed by MCF7 ccl Is, but its binding is specifically inhibited by mannose-6-P and not by other sugars. The presence of mannose-6-P signals indicated that the protein is normally routed to lysosomes where it exerts its usual function (14,15). In testing several enzymatic activities corresponding to lysosomal hydrolases of similar molecular weight, we found that both the purified secreted 52K protein and the corresponding cellular proteins (52K. 34K + 14K) displayed a strong 48K, proteolytic activity at acidic pH, which was mostly inhibited by pepstatin. A comparison to the previously described, and ubiquitous cathepsin-D lead us to conclude to a strong similarity with this lysosomal protease (12). Antibodies to the 52K protein interact with liver cathepsin-D, whi le anti-cathepsin-D immunoprecipitates the 52K protein. The pH and inhibitor sensitivities of the two proteases are very similar, as are their molecular weights. In addition, the first 15 amino acids determined by micro-sequencing the N-terminal of the molecules are identical (P. Ferrara, unpublished).
YIIth International Symposium on Radioimnunology
Inactive NI
f
I
1
2
UK-
pro
cathepsin
379
D
1
1 c
L3 1 autoactivation when
2 +3
intracellular processing hetero activation
secreted 4 51K
4 I
I
I
48K = mature
one
34 K mature
two
chains
3 I
chain
I
I
ILK
Figure 1 Structure end processing of the 52K-cathepsin-D The MCf7-S2K orotein is an inactive oro-cathepsin-D (12.13) which can be autoactivated when ‘secreted or processed’successively into-an intermediate (48h) form and final mature enzyme (34K + 14K). The processed forms of cathepsin-D in normal fibroblasts (23) and in MCF7 breast cancer cells are very similar if not identical. The differences with the previously characterized cathepsin-D of normal tissue linked to its hormonal to be mostly quantitative and at present appear, regulation. The concentration of this enzyme is very high in some breast cancers and its precursor is secreted in greater and melanoma (see clinical studies), mammary epithelial amounts by breast cancer ccl Is (30 to 40%) th an by normal in preparation), cells or fibroblasts in culture (5 to 10%) (F. Capony et al., Until now and to our knowledge, a specific regulation of cathepsin-D by estrogens has never been observed. Using monoclonal antibodies to the 52K protein and synthetic oligonucleotides prepared from the sequence of the N-terminal end of the 52K protein, we screened A gtll (gift of Pr P. a cDNA I i brary expressed in Chambon) and cloned several cDNA probes corresponding to the whole mRNA of the pre-, pro-cathepsin-D of MCF7 ccl Is (16). Partial sequencing indicated a strong homology with cathepsin-D of normal kidney (17) with a few differences whose biological significance is now being studied (P. Augereau et al., in preparation). Using the monoclonal ant ibodies, we showed that the intracellular cathepsin-D was also regulated by using a cDNA probe, that the level of a 2.2. kb 52K pre-, estrogens (18) and, pro-cathepsin-D mRNA was rapidly and specifically increased by 6 to lo-fold It is therefore I ikely that as for other following estradiol treatment (19). estrogen-regulated genes (201, the 52K-cathepsin-D is transcriptionnaly regulated by estrogens. C.
Potential
role
in
manfnary
carcinogenesis
the characteristics of the 52K cathepsin-D, i.e. its high Some of its induction by estrogen, and concentration in proliferative and tumoral cells, the large proportion of its secreted form, suggest that it may have major in function(s) mammary carcinogenesis by stimulating tumor growth and/or The mechanism of the mitogenic invasion via its proteolytic activity (Fig. 2). action of the 52K-cathepsin-D-like enzyme is unknown. As for other proteases, cathepsin-D may acts indirectly by releasing growth factors from precursors or from extracellular matrix via their enzymatic activity and/or by activating proteolytic cleavages are needed to growth factor receptors. For instance, precursor from the membrane (21) and may be involved in the detach the TGFcr activation of TGFR (22).
H. Rochefort et al.
Angiogcnosis Proteoglycans Basement
Prolifora~ion
- Cothopsin - Plasmin.
B
like
membrane
?
activator.
_ ......
Figure 2 Put&iv. functions of estrogen-reguleted secreted proteins, end peptides wcrotod by broaat cantor cells Estrogens, via their nuclear receptors (RE), induce several proteins and factors with are secreted by breast cancer cells. One category (growth factors) Iray act as autocrine factors which stimulate the growth of the same cells (4). Other proteins, such as proteases, can act as mitogens either directly or via the activation of growth factors, and can also facilitate tumor cell invasion. The major normal functions of cathepsins occur in lysosomes at very low pH, where they degrade endogenous proteins. Since a pro-cethepsin-0 enzyme is secreted in large amounts at the periphery of cancer cells, the enzyme may acqu ire abnormal functions, such as facilitating cancer cell migration and invasion by digesting basement membrane, extracellular matrix, and connect i ve tissue. Cathepsin-0 is secreted as an inactive proenzyme, but at acidic pH, it can be euto-act ivated by the removal of a smal I part of the N-terminal pro-fragment (23). The same is true of the breast cancer 52K protein. Culture media conditioned under serum-free conditions by estradiol-treated MCF ccl Is activities that can digest and contain potential proteolytic methemoglo d. In appears to be entirely the 52K proteoglycana. This activity due to pro-cathepsin-D since it is inhibited by pepstatin. It is conceivable that under circumstances, breast cancer cells develop a sufficiently acidic certein microenvironment to allow the autoactivation of the secreted 52K-cathepsin-D. While no direct demonstration of the involvement of cathepsin-II in breast cancer severa I clinical metastasis has yet been obtained, studies performed in and collaboration with several cancer centers parallel, in have strong Iy supported this hypothesis. CLINICAL A.
Timwo
POTENTIAL
distribution
OF THE 52K-CATHEPSIN-D of
imunoperoxidaoe
AS A PROGNOSTIC
TISSUE
MARKER
staining
Using monoclonel antibodies directed to the mature 34K enzyme (20) (Table l), we have examined frozen sect ions of severe I human tissues (24) with the peroxidase-anti-peroxidase technique of Sternberger. Most of the staining was granular in the cytoplasm and corresponded to lysosomes. Among the different normal tissues studied, the protein appeared to be mostly concentrated in sueat glands and liver, but not in norma I uterus or norma I resting mammary glands collected by reduction mammoplesties. By contrast, immunosteining was observed in 43% of 127 biopsies of benign martopathies. Gynecomast i as, fibroadenomas, fibrous leoions, and lobular structures (adenosis, sclerosing adenosis, atypical lobular hyperplasia) were usually negative. The two groups of mastopathies that were highly stained consisted of cysts over 3 mm in diameter and ductal hyperplasias.
Vllth International Symposium on Radioimnunology
Table
381
1. CHARACTERIZATION Specificity
OF THE Aff[nljy n
Site 1 MlG8
MlHll M6HlO D8F5 DllE2 2 D7E3
4;K 3:K
3 M4A3
Site
Site
5
52K
I mmunoprecipitation
PROTEIN lmmunoblot
+ + +
0.83
+
+
0.58
+
+
0.18
+
ND
0.96
+
ND
1:43 1.25 1.00
Site
(KD)
TO THE
+ + +
Fx 52K
YAbs
+ +
+ +
52K only
D9H8
All the monoclonal antibodies (MAbs) are IgGl and purified from a single fusion. The antibodies of sites 1 to 3 (9) recognize the 34K part of the molecule and can be used to assay the tote I 52K-cathepsi n-D in ce I I extracts. The antibodies of sites 4 and 5 recognize only the pro-fragment and can be used (secreted ccl lular) exclusively the precursor alone and of the to assay 52K-cathepsin-D (28). + indicates the specificity of the antigen-antibody recognition. ND : not deternri ned. into proliferative the different histological were pooled When types lesions, (high-risk) and nonproliferative (low-risk) according to criteria we found a significant correlation between defined by DuPont and Page (251, The use of 52K protein staining in proliferation and 52K protein staining. predicting high-risk (proliferative) mastopathies may therefore be useful and since 91.5% of the 52K-protein-negative lesions complementary to histopathology, non-proliferative and 60.5% of the 52K-protein-positive (non cystic) were lesions were proliferative. We also studied 52K-immunostaining of breast cancer tissue in an attempt to its positivity with steroid-receptor content and other prognostic correlate markers. We studied 232 primary breast carcinomas collected from April to in three French cancer centers N,ce (Institut Gustave Roussy, September 1985, 1 . Centre Antoine Lacassagne, Centre Paul Lamarque, Villejuif ; eva I uated’ in frozen Montpel I ier). The 52K protein was sections by concentration of estrogen immunohistochemistry, and the receptors and were assayed in the cytosol by the classical progesterone receptors The 52K protein was detected in 64% of the dextran-coated charcoal method (7). 232 tumor biopsies. The staining was generally less homogeneous in breast cancer cells than in proliferative benign mastopathies and in cells bordering the lumen In the four categories of tumors defined according to their of large cysts. estrogen and progesterone receptors, a similar proportion of positivity for tumors was found to be positive for the 52K protein. No statistical correlation, was observed between the estrogen or progesterone either positive or negative, receptor concentration and the amount of 52K-positive cells. The absence of correlation between these two types of markers was also shown directly using fine-needle aspirates of breast carcinomas to perform double immunohistochemical staining of the nuclear RE and cytoplasmic 52K protein in There were tumor ccl Is with detectable 52K protein and the same sample (26). without RE staining, and others with RE and without detectable 52K protein. other cancers examined some melanomas were found to be positive, while Among, benign nevi were generally negative (24) suggesting that this marker will be useful in melanoma but further studies are required. Histochemistry however, is we therefore completed this study by monitoring difficult to quantify, immunoenrymatic assay of the 52K-cathepsin-0, 6.
Immunoassay
in soluble
extracts
Using two associations of pro-enzyme, both the precursor The technique assayed (Table 1).
ant ibodies that recognize two doma ins of the !52K) and its cellular products (48K + 3410 were was applied both in the culture medium and in
382
H. Rachefort -. et al
to assay the secreted and total cellular forms, respectively. The the cytosol, determined by reference to a standard concentration of the 52K protein was purified 52K protein assayed by silver staining. The product ion and secret ion of the 52K protein were compared in several RE positive and RE negative cell lines cultured in 10% fetal calf serum containing All RE-positive cell lines tested at confluency produced the 52K estrogens. in the presence of estradiol, but its secretion varied markedly protein For instance, according to the cell line and the subspecies. the secret ion of 52K protein by two MCF7 subspecies was quite different, however, its regulation by estrogens was found in all RE positive cell lines (7). Interestingly, the 52K-cathepsin-D was also produced and secreted by the two RE-negat ive breast cancer ce I I I i nes (BT20, MDA-MB231). The secret ion of 52K protein by RE-negative cell lines and their intracellular concentrations were even higher than in some RE-positive cell lines but were not regulated by estrogens. Breast cancer cytoso I is routinely used to determ i ne estrogen and progesterone receptor by radiocompetition and more recently by levels, immunoenzymatic assays using monoclonal antibodies to estrogen and progesterone receptors. Several laboratories and centers have collected banks of frozen cytosol which allow retrospective clinical studies with new markers. We have therefore adapted the immunoassay of the 52K-cathepsin-D to be applied on the breast cancer cytosol routinely used for receptor assays. More precisely, we recently assayed in this cytosol the level of the total cathepsin-D-like protein (using the antibodies to the mature 34K enzyme) (27) and that of the 52K precursor alone (using the antibodies to the pro-fragment) (28). Each assay was a solid phase double-determinant immunoenzymatic assay. Two studies have currently been completed. The first is a prospective study on 183 cytosols of preand post-menopausal primary breast cancers (29) collected in Montpellier from February 1985. The concentration of total cathepsin-D-like enzyme varied markedly according to the patient from 52 to 2,800 U/mg cytosol protein, as did the proportion of the 52K precursor (0 to 42%). The absence of correlation with the estrogen receptor was confirmed (r = 0.41). The correlation with axi I lary invasion for total 52~ concentration lymph node in primary tumors > 700 U/mg protein suggested that the assay of total enzyme in cytosol may be valuable as a prognostic marker for predicting the degree of invasiveness of breast cancers. There was however no correlation with Scarff and Bloom grading, tumor size and age of patients. A second clinical study was retrospective and included on 150 post-menopausal patients of the Danish Breast Cancer Cooperative Group (Dr S. Thorpe and C. Rose), we assayed total 52K enzyme in cytosol collected at surgery F3;: 10 years earlier and kept at -70°C in the Fibiger Center - Finsen Institute . A significant correlation was found between patients with high (‘400 u/mg protein) 52K protein concentrations and short recurrence-free survival. The 52K prote in parameter was found in this study to be independent of the other prognostic parameters. These first clinical studies strongly suggest a role of the 52K-cathepsin-D in the process of tumor invasion and metastasis, which is the major cause of death in breast cancer. However, these resu Its wi I I have to be confirmed and extended in further studies and compared with those of other new prognostic markers (31). CONCLUSIONS
AND PERSPECTIVES
It has long been proposed that proteases are involved in the process of invasion and metastasis by cancer cells (for review see 32). Collagenases (33) and plasminogen activator (34) have been the most extensively studied, and have been considered to be involved in the process of metastasis. At present, there are however few clinical data which show a correlation between protease content in primary tumor and its ability to metastasis, or its overall prognosis. There is however converging evidence indicating that cathepsins are closely related to the metastatic and invasive processes in cancer. Cell transformation and several mitogens have been shown to stimulate the secret ion of the precursors of MEP-cathepsin-L (351, cathepsin-B (36), and the 52K-pro-cathepsin-0. Moreover, the first clinical results indicate that high concentrations of the cathepsin-D-like antigen in the cytosol of primary tumor is associated with invasive breast cancers. These results strongly support the idea that secreted lysosomal proteases as well as other secreted proteases are involved in the process of breast cancer growth and invasion. Cathepsin-D may also be involved in other cancers, but at present our immunoperoxidase staining has revealed high concentrations of this protease mostly in breast cancer cells and melanoma, suggesting that this cathepsin may act more specifically in these two types of cancer, which frequently result in metastases. The involvement of proteases in the contra I of cancer growth has important therapeutic implications. Since the 52K protease, I i ke other autocr ine growth factors, is produced by both hormone-dependent and -independent breast cancer ccl Is, the neutralization ana logs, enzyme etc ) could beof o~sco”,“,“,~e’,“a;l~by antibodies, inhibitors, interest in breast cancer and These treatments would be of more genera I use than melanoma treatments.
VIIth
International
whose efficacy antihormone therapy, Further studi es are needed to prove in carcinogenesis and to discover effects.
Symposiumon Radioinmunology
is restricted the biological inhibit how to
to
383
hormone-dependent cancers. function of these proteases their putative destructive
Acknowledgements We would like to thank members of our laboratory who have contributed to several parts of this work and to E. Barri& and M. EgCa for their skillful preparation of the manuscript. We are gratefu I to Cl in Midy Sanof i Laboratory P. Chambon (Strasbourg) and F. F Paolucci) for monoclonal antibodies, $iUge!E”‘( Pdsteur Paris) for 52K-cDNA cloning, P. Ferrara, P. Louisot, structure studies and to several scientists (Drs M. A. Barrett, for’ protein and the Mason Research Institute) for their gifts Lippman, M. Rich, I. Keydar, Clinical studies have been performed with the help of of mammary ccl I lines. Dr S. in several clinical Centers in Copenhagen (Finsen Institute, Thorpe), Lamarque, F. Laffargue), Villejuif (Drs G. Montpel Iier (Prs H. Pujol, J.L. Delarue) and in Nice (Drs Duplay, Namer). Contesso, This work was supported by the “Institut National de la SantC et de la the “Association Recherche MBdicale”, the Faculty of Medicine of Montpel Iier, pour la Recherche sur le Cancer”, the “FBdBrat ion Nat iona le des Centres de Lutte contre le Cancer” and a grant INSERM-SANOFI n’81039.3. References (1983) Estrogen regulation of and McGuire W.L. 1. Adams D.J., Edwards D.P. specific proteins as a mode of hormone action in human breast cancer. In Biomembranes, vol 11 Rochefort H., Ch:lbbsp;‘389* Garcia M., Veith F., Vignon F. and Capony F., 2. in breast cancer cells in culture : (1984) Effec’i of estrogen Westley B. released proteins and control of cell proliferation. In Hormones and Cancer, and Soto R.J., eds, Alan p. 37, Gurpide E., Calandra R., vol .142, Levy C., R. Liss Inc., New York. (1985) The regulation by estradiol of proteins F. and Rochefort H. 3. Vignon released by breast cancer ccl Is. In Hormone Responsive Tumors, p. 135, Hollander V.P., ed, Academic Press, New York. Huff K., Bates S., Knabbe C., Swain S., M.E., Dickson R.B., 4. Lippman and Gelmann E.P. (1986) Autocrine and McManaway M., Bronzert D., Kasid A. paracrine growth regulation of human breast cancer. Breast Cancer Res. Treat. 1, 59. (1979) Estradiol induced proteins in the MCF7 6. and Rochefort H. 5. Westley human breast cancer cell line. Biochem. Biophys 410’ B. and Rochefort H. induced by 6. Westley (1980) A secre~,“,es,lF~~~~~;eP~~ estrogen in human breast cancer cell lines. Cell 20, 352. M., Duplay H., Cavailles V., Derocq D., Delarue J.C., Contesso G., 7. Garcia Sancho-Garnier H., Richer G., Domergue J., Namer M. and Rochefort Krebs B., distribution of the 52K protein H in mammary tumar(;987) 1mmunohistochemical : A marker associated to ccl I prol iferat ion rather than to hormone responsiveness. J. Steroid Biochem. in press. B., May F.E.B., Brown A.M.C., Krust A., Chambon P., Lippman M.E. and 8. Westley Rochefort H. (1984) Effects of antiestrogens on the estrogen regulated pS2 52-kDa and 16@-kDa protein in MCF7 ccl Is and two tamoxifen resistant RNA, sublines. J. Biol. Chem. 259, 10030. 9 M., Derocq D., Pau B. and Rochefort H. (1985) Simon D., Capony F., , . Garcia Monoclonal antibodies to the estrogen-regulated Mr 52,000 glycoprotein : Characterization and immunodetection in MCF7 cells. Cancer Res. 45, 709. Capdevielle J., Rougeot C., Ferrara P. and Rochefort F., Garcia M., 10. Capony (1986) Purification and characterization of the secreted and cellular H. 52-kDa proteins regulated by estrogens in human breast cancer cells. Eur. J. Biochem. 161, 505. F., Chambon M., 11. Vignon Freiss G., Garcia M. Capony F., and Rochefort H. (1986) Autocrine growth stimulation of the MCF7 breast cancer cells by the estrogen-regulated 52K protein. Endocrinology 118, 1537. F., Morisset M., Barrett A.J., Capony J.P., Broquet P., Vignon F., 12. Capony Chambon M., Louisot P. and (1987) Phosphorylation, Rochefort H. glycosylation and proteolytic activity of the 52K estrogen-induced protein secreted by MCF7 cells. J. Cell. Biol. 104, 253. M., and Rochefort H. Capony F. ( 1986a) Processing and estrogen 13. Morisset regulation of the 52-kDa protein inside MCF7 breast cancer ccl Is. Endocrinology 119, 2773. K. and Hasilik, A. (1985) Lysosomal enzymes and their receptors. 14. Von Figura Ann. Rev. Biochem. 55, 167. A.J. (1970) Purification of isoenzymes from human and chicken liver. 15. Barrett Riochem. J. 117, 601. Augereau P., Garcia M., Cavailles V., Chalbos D., Capony F. and Rochefort H. 16. cDVA and regulation of the messenger cloning RNA for !1987) the estrogen-regulated 52K protease in human breast cancer. In Proceedings of the Endocrine Society Meeting, The Endocrine Society, Bethesda, in press.
384
17. 18.
19.
20.
21.
22.
23.
24.
25. 26.
27.
28.
29.
30. 31.
32.
33.
34. 35. 36
H. Rochefort et al.
(1985) Cloning and sequence and Chirgwin J.M. Faust P.L., Kornfeld S. analysis of cDNA for human cathepsin 0. Proc. Natl. Acad. Sci. USA 82: 4910. Morisset M., Capony F. and Rochefort H. (1986b) The 52-kDa estrogeng;;t;;;d is a lysosomal acidic protease. protein secreted by MCF7 ccl Is Biophys. Res. Commun. 138, 102. and Rochefort H. (1987) Estrogens Cavai I les V., Garcia M. Augereau P., for a pro-cathepsin-D secreted by breast cancer induce the mRNA coding cells. Submitted for publication. Jongstra J., Jakowlew S., Krust A., Chambon P., Dierich A., Gaub M.P., Oudet P. and Reudelhuber T. (1984) Promoter elements of genes Lepennec J.P., coding for proteins and modulation of transcription by estrogens and voI.40, p. 1, Greep progesterone, In Recent Progress in Hormone Research, 0 ed, Academic Press, New York. Chen E.YJ and Goeddel D.V. (1984) DeLynck R., Roberts A.B., Winkler M.E., Human transforming growth factoru: precursor structure and express ion in E. Coli. Cell 38, 287. and Jullien P. (1985) Conversion of a high Lawrence D.A., Pircher R. molecular weight latent R-TGF from chicken embryo fibroblasts into a low Biochem. molecular weight active B-TGF under acidic conditions. Biophys. Res. Commun. 133, 1026. Hasilik A., Von Figura K., Conzelmann E., Nehrkorn H. and Sandhoff K. (1982) of cathepsin D Lysosomal enzyme precursors in human fibroblasts : activation A precursor towards precursor in vitro and activity of B-hexosaminidase. ganglioside GM2. Eur. J. Biochem. 125, 317. Garcia M., Salazar-Retana G., Pages A., Richer G., Domergue J., Pages A.M., Pujol H. and Rochefort H. Cavalie G., Martin J.M., Lamarque J. L., Pau B., (1986) Distribution of the Mr 52,000 estrogen-regulated protein in benign breast diseases and other tissues by immunohistochemistry. Cancer Res. 46, 3734. (1985) Risk factors for breast cancer in women DuPont W.D. and Page D.L. with proliferative breast disease, N. Engl. J. ued. 312, 146. Salazar G., Domergue J., Simony J., Pujol H. and Cavailles V., Garcia M., Rochefort H. (1987) lmmunodetection of estrogen receptor and 52K protein in fine needle aspirates of breast cancer. J. Natl. Cancer Inst. in press. and Pau 8. Paolucci F., Garcia M. Rogier H., Freiss G., (1987) An assay for determining 52K protein in the cytosol of immunoenzymometric breast cancer tissues. Submitted for publication. Rochefort H., Maudelonde t., Cavalie G., Khalaf S. and Vignon F. Freiss G., (1987) Characterization and properties of two monoclonal antibodies to the In Proceedings of the pro-fragment of the 52K estrogen-regulated protease. Endocrine Society, The Endocrine Society, Bethesda, in press. Cavalie G., Freiss G., Benatia M., Maudelonde T., Khalaf S., Frances D., and Rochefort H. Rogier H., Paolucci F., Simony J., Pujol H. (1987) lmmunoenzymatic assay of the estrogen-regulated 52K protease in 183 brerast for parameters. Submitted cancer cytosols. Correlation with other publication. a novel independent prognostic et al. (1987) The 52K-cathepsin-D, Thorpe S. factor in breast cancer. In preparation. Ullrich A. and McGuire W.L. Clark G.M., Wong S.G., Levin W.J., Slamon D.J., and survival with correlation of relapse (1987) Human breast cancer : amplification of the HER-2/neu oncogene. Science 235, 177. (1986) Proteolytic enzymes in tumor invasion and degradation Goldfarb R.H. In Mechanisms of Cancer Metastasis, p. 341, of host extracellular matrices. Powers W.E., Sloane B.F., eds, Martinus Nijhoff Pub., Boston. Honn K.V., Garbisa S., Hart I ., Foltz C.M. and Shafie S. Liotta L.A., Tryggvason K., enzymatic degradation of (1980) Metastatic ootential correlates with basement membrane collagen. Nature 284, 67. inhibit and Reich E. (1983) Antibodies to plasminogen act ivator Ossowski L. human tumor metastasis. Cell 35, 611. of tranformed (1986) The major excreted protein and Gottesman M.M. Gal S. 1760. fibroblasts in an activable acid-protease. J. Biol. Chem. 261, , In Lysosomes (1979) Tumor lysosomal enzymes and invasive growth Poole A.R. eds, Amer i can 304, Dingle J.T. and Fell H.B., in Biology and Pathology, p. Elsevier Publishing Co, New York.
0883-2897/87 s3.00+0.00
1987
Radiat. Appl. Instrum. Part B
Pergamon
Journals Ltd
Printed in Great Britain
THE
ESTROGEN-REGULATED FROM
Henri ROCHEFORT, Franpoise GARCIA, Muriel MORISSET,
52K-CATHEPSIN-D
BIOLOGY
TO CLINICAL
IN
BREAST
CANCER
:
APPLICATIONS
CAPONY, Patrick AUGEREAU, Vincent CAVAILLES, Gilles FREISS, Thierry MAUDELONDE and Franqoise
Unite d’Endocrinologie Cellulaire et Moleculaire Laboratoire de Biologie Cellulaire and Faculte Navacelles, 34100 Montpellier,
de I’INSERM de Mbdecine, France
(U
60
1481, Rue
Marcel VIGNON and de
SUMMARY We have studied estrogen-regulated proteins in an attempt to understand the mechanism by which estrogens stimulate cell proliferation and mammary carcinoIn estrogen receptor positive human breast cancer cell lines (MCF7, genesis. ZR75-1) estrogens specifically increase the production into the culture medium Several high affinity monoclonal of a 52,000 daltons (52K) glycoprotein. antibodies to the partially purified secretory 52K protein have allowed to purify to homogeneity this protein and its cellular processed products. The 52K protein has been identified as the secreted precursor of a cathepsin-D like via signals and routed to I ysosomes protease bearing mannose-6-phosphate The protease is mitogen ic in vitro on estrogen mannose-6-phosphate receptor. deprived MCF7 cells and is able to degrade basement membrane and proteoglycans The cellular related proteins, as detected by immunofollowing its activation. immunoassay are more concentrated in proliferative mammary histochemistry and in breast cancer cytosol ducts than in resting ducts and their concentration lymph nodes invasion and disease free appears to be more correlated with Copenhagen) than with the estrogen receptor (RE) survival (with S. Thorpe, The protein is also produced constitutively by RE-negative ccl I lines, level. variants, it becomes inducible by whi le in some ant iestrogen resistant Cloning of its cDNA in Xgtll contrary to the wild type MCF7 cells. tamoxifen, has a I lowed to show that the mRNA is rapidly induced by estrogens and to it to that of the norma I human kidney sequence the protein and compare cathepsin-D. We conclude that the estrogen-induced cathepsin-D like protease may in stimulating the growth have important autocrine and/or paracrine functions independent breast cancers and may be and invasion of hormone-dependent and useful as a tissue marker to predict high risk mastopathies and breast cancer In addition to other estrogen regulated growth factors, the invasiveness. the I ysosoma I proteases has the potential to stimulate precursor of pro1 iferation and invasiveness of breast cancer cells as long as they are secreted in excess rather than being routed to lysosomes.
INTRODUCTION The development of human breast cancer cell lines has greatly facilitated the study of the hormonal control of cancer cells proliferation and have allowed to define several proteins which are specifically regulated by steroids. Among those, the proteins and peptides which are induced by estrogens before the growth stimulatory effect of estrogens are particularly interesting since some of them may act as mediator of estrogen action to stimulate the growth and/or invasion of breast cancer cells (for review see l-4). We will summarize the studies from our laboratories which have concentrated on one of these proteins, named 52K from its molecular we i ght in denaturing polyacrylamide gel electrophoresis and have lead to its identification and possible function in carcinogenesis. We will then review the clinical potential of this mammary protein which can be localised and assayed in mammary tissue of patients. BIOLOGY
A.
Regulation
and
OF
THE
52K
PROTEIN
OF MCF7
CELLS
purification
?3he 52K protein was first detected by labeling protein of MCF7 cells with 1’ Slmethionine and analysis of the secreted protein by SDS-polyacrylamide gel electrophoresis (5,6). While estradiol increased by 2 to 3 fold the total amount of secreted proteins, normalisation for constant TCA precipitable material showed that some of the proteins (mainly a 160K and a 52K species) were more increased by estradio I from concentrations as low as 1 to 10 specifically picomole.
377
H. Rochefort et al.
378
into the culture medium by secreted in sm II amounts 52K prote in cells/hour) and by other RE-positive esi!zgen-treated MC$ cells (5 ng/lOb ZR75-1). It is constitutively breast cancer ceils under estrogen control (T47D, in RE-negative cell lines produced without requiring the influence of estrogen the protein is specifically (MDA-MB231, BT20) (7). In RE-positive cells, regulated by hormones (estrogens and high doses of androgens) that can bind to and activate the estrogen receptor, but not by glucocorticoids, progestins, or androgens at low concentrations (6). The effects of the antiestrogens, tamoxifen and hydroxytamoxifen, suggested that this protein was in some way related to the mitogenic activity of estrogens. In wild type MCF7 ccl Is, the antiestrogens total ly inhibited the synthesis of the protein, whereas they partially stimulated the synthesis of the progesterone receptor. By contrast, in the antiestrogen-resistant variants of MCF7 cells (R27 and RTx6) cloned for their ability to grow in 1 PM tamoxifen, the antiestrogens became able, I ike estrogens, to increase the production of the 52K protein, but remained unable, in wi Id type MCF7 cells to stimulate the production of the estrogenZIgulated 160K secreted prote:n and of pS2 mRNA (8). In these cell lines, the 52K protein was therefore a better candidate for being a mitogen than the pS2 or 160K proteins. A possible mechasnism for the ant iestrogen resistance was that ccl Is had acquired a growth advantage, due to the tamoxifen induced increased production of autocrine growth factors such as the 52K protein. Using concanavalin-A sepharose chromatography, we partially purified the 52K protein from 22 liters of conditioned medium from MCF7 cell cultures. From this fraction, we obtained several mouse monoclonal antibodies to the 52K protein Final ly, using an immunoaffinity column, we purified the 52K protein to (9). apparent homogeneity (l,OOO-fold purification) both in its secreted and cellular forms (10). The purification of 52K protein made it possible to study its activity on cell growth, determine its structure and identity, and finally clone its corresponding cDNA sequences. By immunopurification without detergent, a homogeneous secreted 52K protein was obta i ned, which was shown to stimulate the growth of estrogen-deprived recipient MCF7 cells (11). This stimulation was dose-dependent and occurred at concentrations (1 to 10 nM) similar to those found in the culture medium. However, time-course experiments indicated that both estradiol nd the 52K prote in required the same (18 h) lag before stimulating I’Hjthymidine incorporation. The in vitro mitogenic activity of the purified 52K protein was in agreement with an estrogen-regulated autocrine mechanism. To understand such mechanism, it is generally necessary to identify the cell surface receptor which is triggered by the mitogen. In our case, we first tried to identify the 52K protein. 6.
Structure
and
identification
as
a cathepsin-D
Study of the coand post-translational modifications of the 52K protein helped us to define i% structure and enzymatic activity. After exposure of cultured MCF7 cells to P, the 52K protein is intensely labeled. However, most of this label can be removed by endoglycosidase-H treatment, which deletes the two N-glycosylated chains of the protein. Mannose-6-P signals were then identified on these chains (12). Pu I se-chase exper i ments and Western b I ot analysis showed that the 52K protein is the precursor of a lysosomal enzyme which accumulates in lysosomes as a stable 34K protein (Fig. 1) (13). About 40% of the ccl lul ar 52K precursor is secreted while about 60% is successively processed into a 48K and a 34K + 14K protein. Part of the secreted 52K protein can be taken up and processed by MCF7 ccl Is, but its binding is specifically inhibited by mannose-6-P and not by other sugars. The presence of mannose-6-P signals indicated that the protein is normally routed to lysosomes where it exerts its usual function (14,15). In testing several enzymatic activities corresponding to lysosomal hydrolases of similar molecular weight, we found that both the purified secreted 52K protein and the corresponding cellular proteins (52K. 34K + 14K) displayed a strong 48K, proteolytic activity at acidic pH, which was mostly inhibited by pepstatin. A comparison to the previously described, and ubiquitous cathepsin-D lead us to conclude to a strong similarity with this lysosomal protease (12). Antibodies to the 52K protein interact with liver cathepsin-D, whi le anti-cathepsin-D immunoprecipitates the 52K protein. The pH and inhibitor sensitivities of the two proteases are very similar, as are their molecular weights. In addition, the first 15 amino acids determined by micro-sequencing the N-terminal of the molecules are identical (P. Ferrara, unpublished).
YIIth International Symposium on Radioimnunology
Inactive NI
f
I
1
2
UK-
pro
cathepsin
379
D
1
1 c
L3 1 autoactivation when
2 +3
intracellular processing hetero activation
secreted 4 51K
4 I
I
I
48K = mature
one
34 K mature
two
chains
3 I
chain
I
I
ILK
Figure 1 Structure end processing of the 52K-cathepsin-D The MCf7-S2K orotein is an inactive oro-cathepsin-D (12.13) which can be autoactivated when ‘secreted or processed’successively into-an intermediate (48h) form and final mature enzyme (34K + 14K). The processed forms of cathepsin-D in normal fibroblasts (23) and in MCF7 breast cancer cells are very similar if not identical. The differences with the previously characterized cathepsin-D of normal tissue linked to its hormonal to be mostly quantitative and at present appear, regulation. The concentration of this enzyme is very high in some breast cancers and its precursor is secreted in greater and melanoma (see clinical studies), mammary epithelial amounts by breast cancer ccl Is (30 to 40%) th an by normal in preparation), cells or fibroblasts in culture (5 to 10%) (F. Capony et al., Until now and to our knowledge, a specific regulation of cathepsin-D by estrogens has never been observed. Using monoclonal antibodies to the 52K protein and synthetic oligonucleotides prepared from the sequence of the N-terminal end of the 52K protein, we screened A gtll (gift of Pr P. a cDNA I i brary expressed in Chambon) and cloned several cDNA probes corresponding to the whole mRNA of the pre-, pro-cathepsin-D of MCF7 ccl Is (16). Partial sequencing indicated a strong homology with cathepsin-D of normal kidney (17) with a few differences whose biological significance is now being studied (P. Augereau et al., in preparation). Using the monoclonal ant ibodies, we showed that the intracellular cathepsin-D was also regulated by using a cDNA probe, that the level of a 2.2. kb 52K pre-, estrogens (18) and, pro-cathepsin-D mRNA was rapidly and specifically increased by 6 to lo-fold It is therefore I ikely that as for other following estradiol treatment (19). estrogen-regulated genes (201, the 52K-cathepsin-D is transcriptionnaly regulated by estrogens. C.
Potential
role
in
manfnary
carcinogenesis
the characteristics of the 52K cathepsin-D, i.e. its high Some of its induction by estrogen, and concentration in proliferative and tumoral cells, the large proportion of its secreted form, suggest that it may have major in function(s) mammary carcinogenesis by stimulating tumor growth and/or The mechanism of the mitogenic invasion via its proteolytic activity (Fig. 2). action of the 52K-cathepsin-D-like enzyme is unknown. As for other proteases, cathepsin-D may acts indirectly by releasing growth factors from precursors or from extracellular matrix via their enzymatic activity and/or by activating proteolytic cleavages are needed to growth factor receptors. For instance, precursor from the membrane (21) and may be involved in the detach the TGFcr activation of TGFR (22).
H. Rochefort et al.
Angiogcnosis Proteoglycans Basement
Prolifora~ion
- Cothopsin - Plasmin.
B
like
membrane
?
activator.
_ ......
Figure 2 Put&iv. functions of estrogen-reguleted secreted proteins, end peptides wcrotod by broaat cantor cells Estrogens, via their nuclear receptors (RE), induce several proteins and factors with are secreted by breast cancer cells. One category (growth factors) Iray act as autocrine factors which stimulate the growth of the same cells (4). Other proteins, such as proteases, can act as mitogens either directly or via the activation of growth factors, and can also facilitate tumor cell invasion. The major normal functions of cathepsins occur in lysosomes at very low pH, where they degrade endogenous proteins. Since a pro-cethepsin-0 enzyme is secreted in large amounts at the periphery of cancer cells, the enzyme may acqu ire abnormal functions, such as facilitating cancer cell migration and invasion by digesting basement membrane, extracellular matrix, and connect i ve tissue. Cathepsin-0 is secreted as an inactive proenzyme, but at acidic pH, it can be euto-act ivated by the removal of a smal I part of the N-terminal pro-fragment (23). The same is true of the breast cancer 52K protein. Culture media conditioned under serum-free conditions by estradiol-treated MCF ccl Is activities that can digest and contain potential proteolytic methemoglo d. In appears to be entirely the 52K proteoglycana. This activity due to pro-cathepsin-D since it is inhibited by pepstatin. It is conceivable that under circumstances, breast cancer cells develop a sufficiently acidic certein microenvironment to allow the autoactivation of the secreted 52K-cathepsin-D. While no direct demonstration of the involvement of cathepsin-II in breast cancer severa I clinical metastasis has yet been obtained, studies performed in and collaboration with several cancer centers parallel, in have strong Iy supported this hypothesis. CLINICAL A.
Timwo
POTENTIAL
distribution
OF THE 52K-CATHEPSIN-D of
imunoperoxidaoe
AS A PROGNOSTIC
TISSUE
MARKER
staining
Using monoclonel antibodies directed to the mature 34K enzyme (20) (Table l), we have examined frozen sect ions of severe I human tissues (24) with the peroxidase-anti-peroxidase technique of Sternberger. Most of the staining was granular in the cytoplasm and corresponded to lysosomes. Among the different normal tissues studied, the protein appeared to be mostly concentrated in sueat glands and liver, but not in norma I uterus or norma I resting mammary glands collected by reduction mammoplesties. By contrast, immunosteining was observed in 43% of 127 biopsies of benign martopathies. Gynecomast i as, fibroadenomas, fibrous leoions, and lobular structures (adenosis, sclerosing adenosis, atypical lobular hyperplasia) were usually negative. The two groups of mastopathies that were highly stained consisted of cysts over 3 mm in diameter and ductal hyperplasias.
Vllth International Symposium on Radioimnunology
Table
381
1. CHARACTERIZATION Specificity
OF THE Aff[nljy n
Site 1 MlG8
MlHll M6HlO D8F5 DllE2 2 D7E3
4;K 3:K
3 M4A3
Site
Site
5
52K
I mmunoprecipitation
PROTEIN lmmunoblot
+ + +
0.83
+
+
0.58
+
+
0.18
+
ND
0.96
+
ND
1:43 1.25 1.00
Site
(KD)
TO THE
+ + +
Fx 52K
YAbs
+ +
+ +
52K only
D9H8
All the monoclonal antibodies (MAbs) are IgGl and purified from a single fusion. The antibodies of sites 1 to 3 (9) recognize the 34K part of the molecule and can be used to assay the tote I 52K-cathepsi n-D in ce I I extracts. The antibodies of sites 4 and 5 recognize only the pro-fragment and can be used (secreted ccl lular) exclusively the precursor alone and of the to assay 52K-cathepsin-D (28). + indicates the specificity of the antigen-antibody recognition. ND : not deternri ned. into proliferative the different histological were pooled When types lesions, (high-risk) and nonproliferative (low-risk) according to criteria we found a significant correlation between defined by DuPont and Page (251, The use of 52K protein staining in proliferation and 52K protein staining. predicting high-risk (proliferative) mastopathies may therefore be useful and since 91.5% of the 52K-protein-negative lesions complementary to histopathology, non-proliferative and 60.5% of the 52K-protein-positive (non cystic) were lesions were proliferative. We also studied 52K-immunostaining of breast cancer tissue in an attempt to its positivity with steroid-receptor content and other prognostic correlate markers. We studied 232 primary breast carcinomas collected from April to in three French cancer centers N,ce (Institut Gustave Roussy, September 1985, 1 . Centre Antoine Lacassagne, Centre Paul Lamarque, Villejuif ; eva I uated’ in frozen Montpel I ier). The 52K protein was sections by concentration of estrogen immunohistochemistry, and the receptors and were assayed in the cytosol by the classical progesterone receptors The 52K protein was detected in 64% of the dextran-coated charcoal method (7). 232 tumor biopsies. The staining was generally less homogeneous in breast cancer cells than in proliferative benign mastopathies and in cells bordering the lumen In the four categories of tumors defined according to their of large cysts. estrogen and progesterone receptors, a similar proportion of positivity for tumors was found to be positive for the 52K protein. No statistical correlation, was observed between the estrogen or progesterone either positive or negative, receptor concentration and the amount of 52K-positive cells. The absence of correlation between these two types of markers was also shown directly using fine-needle aspirates of breast carcinomas to perform double immunohistochemical staining of the nuclear RE and cytoplasmic 52K protein in There were tumor ccl Is with detectable 52K protein and the same sample (26). without RE staining, and others with RE and without detectable 52K protein. other cancers examined some melanomas were found to be positive, while Among, benign nevi were generally negative (24) suggesting that this marker will be useful in melanoma but further studies are required. Histochemistry however, is we therefore completed this study by monitoring difficult to quantify, immunoenrymatic assay of the 52K-cathepsin-0, 6.
Immunoassay
in soluble
extracts
Using two associations of pro-enzyme, both the precursor The technique assayed (Table 1).
ant ibodies that recognize two doma ins of the !52K) and its cellular products (48K + 3410 were was applied both in the culture medium and in
382
H. Rachefort -. et al
to assay the secreted and total cellular forms, respectively. The the cytosol, determined by reference to a standard concentration of the 52K protein was purified 52K protein assayed by silver staining. The product ion and secret ion of the 52K protein were compared in several RE positive and RE negative cell lines cultured in 10% fetal calf serum containing All RE-positive cell lines tested at confluency produced the 52K estrogens. in the presence of estradiol, but its secretion varied markedly protein For instance, according to the cell line and the subspecies. the secret ion of 52K protein by two MCF7 subspecies was quite different, however, its regulation by estrogens was found in all RE positive cell lines (7). Interestingly, the 52K-cathepsin-D was also produced and secreted by the two RE-negat ive breast cancer ce I I I i nes (BT20, MDA-MB231). The secret ion of 52K protein by RE-negative cell lines and their intracellular concentrations were even higher than in some RE-positive cell lines but were not regulated by estrogens. Breast cancer cytoso I is routinely used to determ i ne estrogen and progesterone receptor by radiocompetition and more recently by levels, immunoenzymatic assays using monoclonal antibodies to estrogen and progesterone receptors. Several laboratories and centers have collected banks of frozen cytosol which allow retrospective clinical studies with new markers. We have therefore adapted the immunoassay of the 52K-cathepsin-D to be applied on the breast cancer cytosol routinely used for receptor assays. More precisely, we recently assayed in this cytosol the level of the total cathepsin-D-like protein (using the antibodies to the mature 34K enzyme) (27) and that of the 52K precursor alone (using the antibodies to the pro-fragment) (28). Each assay was a solid phase double-determinant immunoenzymatic assay. Two studies have currently been completed. The first is a prospective study on 183 cytosols of preand post-menopausal primary breast cancers (29) collected in Montpellier from February 1985. The concentration of total cathepsin-D-like enzyme varied markedly according to the patient from 52 to 2,800 U/mg cytosol protein, as did the proportion of the 52K precursor (0 to 42%). The absence of correlation with the estrogen receptor was confirmed (r = 0.41). The correlation with axi I lary invasion for total 52~ concentration lymph node in primary tumors > 700 U/mg protein suggested that the assay of total enzyme in cytosol may be valuable as a prognostic marker for predicting the degree of invasiveness of breast cancers. There was however no correlation with Scarff and Bloom grading, tumor size and age of patients. A second clinical study was retrospective and included on 150 post-menopausal patients of the Danish Breast Cancer Cooperative Group (Dr S. Thorpe and C. Rose), we assayed total 52K enzyme in cytosol collected at surgery F3;: 10 years earlier and kept at -70°C in the Fibiger Center - Finsen Institute . A significant correlation was found between patients with high (‘400 u/mg protein) 52K protein concentrations and short recurrence-free survival. The 52K prote in parameter was found in this study to be independent of the other prognostic parameters. These first clinical studies strongly suggest a role of the 52K-cathepsin-D in the process of tumor invasion and metastasis, which is the major cause of death in breast cancer. However, these resu Its wi I I have to be confirmed and extended in further studies and compared with those of other new prognostic markers (31). CONCLUSIONS
AND PERSPECTIVES
It has long been proposed that proteases are involved in the process of invasion and metastasis by cancer cells (for review see 32). Collagenases (33) and plasminogen activator (34) have been the most extensively studied, and have been considered to be involved in the process of metastasis. At present, there are however few clinical data which show a correlation between protease content in primary tumor and its ability to metastasis, or its overall prognosis. There is however converging evidence indicating that cathepsins are closely related to the metastatic and invasive processes in cancer. Cell transformation and several mitogens have been shown to stimulate the secret ion of the precursors of MEP-cathepsin-L (351, cathepsin-B (36), and the 52K-pro-cathepsin-0. Moreover, the first clinical results indicate that high concentrations of the cathepsin-D-like antigen in the cytosol of primary tumor is associated with invasive breast cancers. These results strongly support the idea that secreted lysosomal proteases as well as other secreted proteases are involved in the process of breast cancer growth and invasion. Cathepsin-D may also be involved in other cancers, but at present our immunoperoxidase staining has revealed high concentrations of this protease mostly in breast cancer cells and melanoma, suggesting that this cathepsin may act more specifically in these two types of cancer, which frequently result in metastases. The involvement of proteases in the contra I of cancer growth has important therapeutic implications. Since the 52K protease, I i ke other autocr ine growth factors, is produced by both hormone-dependent and -independent breast cancer ccl Is, the neutralization ana logs, enzyme etc ) could beof o~sco”,“,“,~e’,“a;l~by antibodies, inhibitors, interest in breast cancer and These treatments would be of more genera I use than melanoma treatments.
VIIth
International
whose efficacy antihormone therapy, Further studi es are needed to prove in carcinogenesis and to discover effects.
Symposiumon Radioinmunology
is restricted the biological inhibit how to
to
383
hormone-dependent cancers. function of these proteases their putative destructive
Acknowledgements We would like to thank members of our laboratory who have contributed to several parts of this work and to E. Barri& and M. EgCa for their skillful preparation of the manuscript. We are gratefu I to Cl in Midy Sanof i Laboratory P. Chambon (Strasbourg) and F. F Paolucci) for monoclonal antibodies, $iUge!E”‘( Pdsteur Paris) for 52K-cDNA cloning, P. Ferrara, P. Louisot, structure studies and to several scientists (Drs M. A. Barrett, for’ protein and the Mason Research Institute) for their gifts Lippman, M. Rich, I. Keydar, Clinical studies have been performed with the help of of mammary ccl I lines. Dr S. in several clinical Centers in Copenhagen (Finsen Institute, Thorpe), Lamarque, F. Laffargue), Villejuif (Drs G. Montpel Iier (Prs H. Pujol, J.L. Delarue) and in Nice (Drs Duplay, Namer). Contesso, This work was supported by the “Institut National de la SantC et de la the “Association Recherche MBdicale”, the Faculty of Medicine of Montpel Iier, pour la Recherche sur le Cancer”, the “FBdBrat ion Nat iona le des Centres de Lutte contre le Cancer” and a grant INSERM-SANOFI n’81039.3. References (1983) Estrogen regulation of and McGuire W.L. 1. Adams D.J., Edwards D.P. specific proteins as a mode of hormone action in human breast cancer. In Biomembranes, vol 11 Rochefort H., Ch:lbbsp;‘389* Garcia M., Veith F., Vignon F. and Capony F., 2. in breast cancer cells in culture : (1984) Effec’i of estrogen Westley B. released proteins and control of cell proliferation. In Hormones and Cancer, and Soto R.J., eds, Alan p. 37, Gurpide E., Calandra R., vol .142, Levy C., R. Liss Inc., New York. (1985) The regulation by estradiol of proteins F. and Rochefort H. 3. Vignon released by breast cancer ccl Is. In Hormone Responsive Tumors, p. 135, Hollander V.P., ed, Academic Press, New York. Huff K., Bates S., Knabbe C., Swain S., M.E., Dickson R.B., 4. Lippman and Gelmann E.P. (1986) Autocrine and McManaway M., Bronzert D., Kasid A. paracrine growth regulation of human breast cancer. Breast Cancer Res. Treat. 1, 59. (1979) Estradiol induced proteins in the MCF7 6. and Rochefort H. 5. Westley human breast cancer cell line. Biochem. Biophys 410’ B. and Rochefort H. induced by 6. Westley (1980) A secre~,“,es,lF~~~~~;eP~~ estrogen in human breast cancer cell lines. Cell 20, 352. M., Duplay H., Cavailles V., Derocq D., Delarue J.C., Contesso G., 7. Garcia Sancho-Garnier H., Richer G., Domergue J., Namer M. and Rochefort Krebs B., distribution of the 52K protein H in mammary tumar(;987) 1mmunohistochemical : A marker associated to ccl I prol iferat ion rather than to hormone responsiveness. J. Steroid Biochem. in press. B., May F.E.B., Brown A.M.C., Krust A., Chambon P., Lippman M.E. and 8. Westley Rochefort H. (1984) Effects of antiestrogens on the estrogen regulated pS2 52-kDa and 16@-kDa protein in MCF7 ccl Is and two tamoxifen resistant RNA, sublines. J. Biol. Chem. 259, 10030. 9 M., Derocq D., Pau B. and Rochefort H. (1985) Simon D., Capony F., , . Garcia Monoclonal antibodies to the estrogen-regulated Mr 52,000 glycoprotein : Characterization and immunodetection in MCF7 cells. Cancer Res. 45, 709. Capdevielle J., Rougeot C., Ferrara P. and Rochefort F., Garcia M., 10. Capony (1986) Purification and characterization of the secreted and cellular H. 52-kDa proteins regulated by estrogens in human breast cancer cells. Eur. J. Biochem. 161, 505. F., Chambon M., 11. Vignon Freiss G., Garcia M. Capony F., and Rochefort H. (1986) Autocrine growth stimulation of the MCF7 breast cancer cells by the estrogen-regulated 52K protein. Endocrinology 118, 1537. F., Morisset M., Barrett A.J., Capony J.P., Broquet P., Vignon F., 12. Capony Chambon M., Louisot P. and (1987) Phosphorylation, Rochefort H. glycosylation and proteolytic activity of the 52K estrogen-induced protein secreted by MCF7 cells. J. Cell. Biol. 104, 253. M., and Rochefort H. Capony F. ( 1986a) Processing and estrogen 13. Morisset regulation of the 52-kDa protein inside MCF7 breast cancer ccl Is. Endocrinology 119, 2773. K. and Hasilik, A. (1985) Lysosomal enzymes and their receptors. 14. Von Figura Ann. Rev. Biochem. 55, 167. A.J. (1970) Purification of isoenzymes from human and chicken liver. 15. Barrett Riochem. J. 117, 601. Augereau P., Garcia M., Cavailles V., Chalbos D., Capony F. and Rochefort H. 16. cDVA and regulation of the messenger cloning RNA for !1987) the estrogen-regulated 52K protease in human breast cancer. In Proceedings of the Endocrine Society Meeting, The Endocrine Society, Bethesda, in press.
384
17. 18.
19.
20.
21.
22.
23.
24.
25. 26.
27.
28.
29.
30. 31.
32.
33.
34. 35. 36
H. Rochefort et al.
(1985) Cloning and sequence and Chirgwin J.M. Faust P.L., Kornfeld S. analysis of cDNA for human cathepsin 0. Proc. Natl. Acad. Sci. USA 82: 4910. Morisset M., Capony F. and Rochefort H. (1986b) The 52-kDa estrogeng;;t;;;d is a lysosomal acidic protease. protein secreted by MCF7 ccl Is Biophys. Res. Commun. 138, 102. and Rochefort H. (1987) Estrogens Cavai I les V., Garcia M. Augereau P., for a pro-cathepsin-D secreted by breast cancer induce the mRNA coding cells. Submitted for publication. Jongstra J., Jakowlew S., Krust A., Chambon P., Dierich A., Gaub M.P., Oudet P. and Reudelhuber T. (1984) Promoter elements of genes Lepennec J.P., coding for proteins and modulation of transcription by estrogens and voI.40, p. 1, Greep progesterone, In Recent Progress in Hormone Research, 0 ed, Academic Press, New York. Chen E.YJ and Goeddel D.V. (1984) DeLynck R., Roberts A.B., Winkler M.E., Human transforming growth factoru: precursor structure and express ion in E. Coli. Cell 38, 287. and Jullien P. (1985) Conversion of a high Lawrence D.A., Pircher R. molecular weight latent R-TGF from chicken embryo fibroblasts into a low Biochem. molecular weight active B-TGF under acidic conditions. Biophys. Res. Commun. 133, 1026. Hasilik A., Von Figura K., Conzelmann E., Nehrkorn H. and Sandhoff K. (1982) of cathepsin D Lysosomal enzyme precursors in human fibroblasts : activation A precursor towards precursor in vitro and activity of B-hexosaminidase. ganglioside GM2. Eur. J. Biochem. 125, 317. Garcia M., Salazar-Retana G., Pages A., Richer G., Domergue J., Pages A.M., Pujol H. and Rochefort H. Cavalie G., Martin J.M., Lamarque J. L., Pau B., (1986) Distribution of the Mr 52,000 estrogen-regulated protein in benign breast diseases and other tissues by immunohistochemistry. Cancer Res. 46, 3734. (1985) Risk factors for breast cancer in women DuPont W.D. and Page D.L. with proliferative breast disease, N. Engl. J. ued. 312, 146. Salazar G., Domergue J., Simony J., Pujol H. and Cavailles V., Garcia M., Rochefort H. (1987) lmmunodetection of estrogen receptor and 52K protein in fine needle aspirates of breast cancer. J. Natl. Cancer Inst. in press. and Pau 8. Paolucci F., Garcia M. Rogier H., Freiss G., (1987) An assay for determining 52K protein in the cytosol of immunoenzymometric breast cancer tissues. Submitted for publication. Rochefort H., Maudelonde t., Cavalie G., Khalaf S. and Vignon F. Freiss G., (1987) Characterization and properties of two monoclonal antibodies to the In Proceedings of the pro-fragment of the 52K estrogen-regulated protease. Endocrine Society, The Endocrine Society, Bethesda, in press. Cavalie G., Freiss G., Benatia M., Maudelonde T., Khalaf S., Frances D., and Rochefort H. Rogier H., Paolucci F., Simony J., Pujol H. (1987) lmmunoenzymatic assay of the estrogen-regulated 52K protease in 183 brerast for parameters. Submitted cancer cytosols. Correlation with other publication. a novel independent prognostic et al. (1987) The 52K-cathepsin-D, Thorpe S. factor in breast cancer. In preparation. Ullrich A. and McGuire W.L. Clark G.M., Wong S.G., Levin W.J., Slamon D.J., and survival with correlation of relapse (1987) Human breast cancer : amplification of the HER-2/neu oncogene. Science 235, 177. (1986) Proteolytic enzymes in tumor invasion and degradation Goldfarb R.H. In Mechanisms of Cancer Metastasis, p. 341, of host extracellular matrices. Powers W.E., Sloane B.F., eds, Martinus Nijhoff Pub., Boston. Honn K.V., Garbisa S., Hart I ., Foltz C.M. and Shafie S. Liotta L.A., Tryggvason K., enzymatic degradation of (1980) Metastatic ootential correlates with basement membrane collagen. Nature 284, 67. inhibit and Reich E. (1983) Antibodies to plasminogen act ivator Ossowski L. human tumor metastasis. Cell 35, 611. of tranformed (1986) The major excreted protein and Gottesman M.M. Gal S. 1760. fibroblasts in an activable acid-protease. J. Biol. Chem. 261, , In Lysosomes (1979) Tumor lysosomal enzymes and invasive growth Poole A.R. eds, Amer i can 304, Dingle J.T. and Fell H.B., in Biology and Pathology, p. Elsevier Publishing Co, New York.