Extracellular matrix proteins and their receptors in the normal, hyperplastic and neoplastic breast

Cell Differentiation and Development, 32 (1990) 409-416 © 1990 Elsevier Scientific Publishers Ireland, Ltd. 0922-3371/90/$03.50

409

CELDIF 99926

Extracellular matrix proteins and their receptors in the normal, hyperplastic and neoplastic breast Victor E. G o u l d

1, G e o r g e

K. K o u k o u l i s 1 a n d I s m o V i r t a n e n 2

l Department of Pathology, Rush Medical College, Chicago, Illinois, U.S.A. and 2 Department of Anatomy, University of Helsinki, Helsinki, Finland

We studied by immunohistochemistry, the distribution of tenascin (Ten), cellular fibronectin (cFn), iaminin and certain pertinent extracellular matrix protein receptors in normal human female breast, variants of fibrocystic disease (FCD), benign tumors, and ductal and Iobular carcinomas. Monocional antibodies (mAb) to Ten, extradomain A containing cFn (EDAcFn), A and B chains of iaminin, and beta-1 (13-1) and different a subunits of integrins were used. In in-situ ductal and Iobular carcinomas, iaminin staining had focal gaps, Ten-immunoreactivity displayed periductal or periacinar bands, and cFn showed broad and intense periductal staining; strong reactions for [3-1 and a-6 were noted in the basal cytoplasm of non-neoplastic myoepithelial cells while few tumor cells stained weakly. In infiltrating ductal and lobular carcinomas (IDC, ILC), iaminin reactivity was weak, uneven or absent around neoplastic clusters whereas stromal staining for Ten and cFn was extensive and strong. In most IDC, moderate 13-1 and c~-6 staining involved variable subpopulations; one mucinous carcinoma stained strongly and diffusely. In 20-40% of cells in ILC, [3-1 and ct-6 were localized in delicate, ramified cytoplasmic processes. Indirect immunofluorescence studies with mAbs to other ct-integrin subunits suggest that in various breast carcinomas only ct-3 is expressed in tumor cells and that the vessels contained a-1 integrin. As compared with the normal breast, FCD and benign tumors, reactivity for Ten and cFn is increased in breast carcinomas while laminin is attenuated and decreased or absent; yet, Ten cannot be regarded as a carcinoma marker since it can be detected in benign tumors, FCD, and even in the normal breast. Reactivity for 13-1 and et-6 integrin subunits is decreased in all breast carcinomas; however, the comparatively strong and distinctly localized reactions noted in lobular vs ductal carcinomas may well reflect biological and clinical differences between these two main breast carcinoma phenotypes. Normal breast; Breast tumor; Tenascin; Cellular fibronectin; Laminin; Integrin; Immunohistocbemistry

Introduction Interaction phenomena between mesenchymal and epithelial - or other parenchymal - compoCorrespondence address: V.E. Gould, Department of Pathology, Rush Medical College, 1653 West Congress Pkwy, Chicago, IL 60612-3684, U.S.A.

nents are fundamental for normal embryogenesis of all organs and tissues including the breast (Kratochwil, 1969; Ekblom et al., 1990). These phenomena remain important for the maintenance of normal structures in adults as well as for reparative, hyperplastic and neoplastic processes (Sakakura, 1983). It has been suggested that certain aspects of mesenchymal induction of epi-

410

thelial development could be mediated by extracellular matrix (ECM) molecules (Grobstein, 1954) that include proteoglycans, collagens and non-collagenous glycoproteins (for overviews see Hay, 1989; Liotta, 1984). In this context, fibronectin and laminin have been said to play important roles in regulating cell migration, morphogenesis and differentiation (MartinezHernandez, 1988; Martinez-Hernandez and Amenta, 1983). Tenascin (Ten) is a rather recently isolated and characterized ECM glycoprotein named by Chiquet-Ehrismann et al. (1986). Early studies documented the presence and possible role of Ten during cartilage and bone development (for reviews, see Ekblom, 1990; Chiquet-Ehrismann, 1990). Notably, Ten was thought to decrease after birth to the extent that it could not be detected in the adult rat breast (Inaguma et al., 1988; Mackie et al., 1987a) while in the adult murine kidney Ten persisted only focally in the medulla but not in the cortex (Aufderheide et al., 1987). On the other hand, immunohistochemical and biochemical studies showed Ten in rat, mouse and human mammary carcinomas while it could not be shown in benign proliferations of the same organ (Chiquet-Ehrismann et al., 1986; Inaguma et al., 1988; Mackie et al., 1987a). Thus, Ten seemed to share features shown by several oncodevelopmental-type molecules including its rapid re-emergence during reparative phenomena as exemplified by its abundance at the epithelial-stromal interface and in the granulation tissue of experimentally induced skin wounds (Mackie et al., 1988a). As a part of a broader program (Koukoulis et

al., 1991), we undertook a systematic immunohistochemical and biochemical study on the distribution of Ten in snap-frozen normal and pathologic human breast samples with the recently characterized mAb 100EB2 (Howeedy et al., 1990). In this brief overview, we will outline our observations and attempt to correlate them with the findings obtained in the same samples with mAbs to EDAcFn, A and B chains of laminin, and the fl-1 and a-6 subunits of integrins.

Materials and Methods

Ninety-two snap frozen breast samples were studied including normal adult female (n = 8), fetal female (n = 4), gestational (n--4), elderly female without evidence of fibrocystic disease (FCD) (n = 5), FCD (n = 11, including variable admixtures of fibrosis, large cysts, adenosis, apocrine metaplasia, ductal hyperplasia and papillomas), fibroadenomas (n = 3), cystosarcoma phylloides (n = 2), ductal and lobular carcinomas (n = 26 and n = 27 respectively; in-situ loci were included in several samples). Most of these cases have been extensively studied (Gould et al., 1988, 1990; Koukoulis et al., 1991). Monoclonal antibodies used were: (a) 100EB2 to Ten recognizing hexabrachion polypeptides of M r 180,000 and 250,000 (Howeedy et al., 1990), (b) 52DH1 to an extra domain A of cFn distinct from plasma fibronectin (Vartio et al., 1987), (c) and (d) 113BC7 and ll4DG10 to the A and B chains of laminin derived from placental extracts (Wewer et al., 1983) and reacting with M r 400,000

Fig. 1. (a). Normal adult breast immunostalned for Ten; note immunoreactive rim encompassing lobule as well as individual acini while the surrounding stroma is negative. Original magnification x260. (b). Normal adult breast immunostained for EDAcFn; observe reactivity encompassing periductal fibroblasts while the space between them and the myoepithelial layer is negative (arrow). Original magnification x 380. (c). Normal adult resting breast including terminal ducts and parts of lobule immunostained for the fl-1 subunit of Int.. Immunoreactivity involves predominantly the basal aspect of the myoepithelial cells (arrows) while faint staining is noted in the basolateral aspects of some lamina] cells (arrowhead). Original magnification × 380. (d). FCD; segment of florid but benign papilloma immunostalned for Ten. Note virtually uninterrupted immunoreaction along the thin fibrovascular septa. Original magnification × 260. (e). Intraductal (in-situ) carcinoma immunostained for the a-6 subunit of Int.; convincing reactiyity restricted to remaining, non-neoplastic focally discontinuous layer of myoepithelial cells (arrows). Original magnification x 460. (f). Infiltrating ductal carcinoma not otherwise specified immunostained for the fl-1 subunit of Int.; a majority of cells show faint but convincing staining. Original magnification x 460. (g). Infiltrating ductal carcinoma, mucinous variant, immunostained for the fl-1 subunit of Int.. Note intense reaction restricted to the cell aggregate in the non-reactive mucus pool. Original magnification x460. (h). Infiltrating lobular carcinoma immunostalned for the a-6 subunit of Int.. Observe reactivity in the notable but delicate cytoplastic processes (arrows). Original magnification x 660.

411

412 and 200,000 polypeptides, respectively, (e) 102DF5 to the fl-1 subunit of integrins (Yl~inne et al., 1989), and (f) Ts 2/7, 10Gll, JI43, BIE5 and GoH3 to the a-l, a-2, a-3, a-5 and a-6 subunits of integrins, respectively (Sonnenberg et al., 1987). Cold acetone fixed, 4/~m thick cryosections were stained by the ABC method (Hsu et al., 1981). Binding sites were visualized with 3,3'-diaminobenzidine; the slides were briefly counterstained with hematoxylin. Selected samples were studied by double immunofluorescence microscopy with the mAb 100EB2 and a well characterized rabbit antiserum to laminin (Liesi et al., 1983). mAbs to a-l, a-2, a-4 and a-5 subunits of integrins were applied in a preliminary survey by using the indirect immunofluorescence method.

Results

Normal breast During the reproductive years and in elderly females (without FCD), Ten was consistently found as a rather thin periductal and periacinar stromal rim (Fig. la). Reactions for EDAcFn appeared encompassing fibroblast-like cells around ducts and less so around acini; cFn staining was not evident in the immediate vicinity of the myoepithelium nor did it extend broadly into the interductal stroma (Fig. lb). Staining with laminin antibodies showed consistently single and thin periductal and periacinar linear reactions in the basement membrane region following closely the contour of the myoepithelial layer. In all breast samples studied, the reaction was similar with all laminin antibodies applied. Immunoreactivity for fl-1 and a-6 subunit of integrins was localized predominantly in the basal cytoplasm and in cytoplasmic processes of myoepithelial cells (Fig. lc); weak fl-1 staining was also seen in the basal and basolateral aspects of luminal cells of ducts and acini and appeared to correspond to reactivity obtained with mAbs to a-2 and a-3. These distribution patterns remained unchanged in samples of fetal and gestational breast except for notable increases in the extent and intensity of Ten and cFn staining reactions; the latter was particularly diffuse and focally "raggedy" in fetal samples.

FCD Variably altered ductal and acinar structures were encompassed by a delicate line of laminin reactivity; in rare foci, double and even multiple layers of laminin were noted. Ten and cFn were variably increased in the stroma surrounding dilated, metaplastic a n d / o r hyperplastic ducts and acini and in papillary lesions of all types (Fig. ld). For the most part, the reactions were limited to the stroma surrounding the epithelial structures and did not extend into the distal, densely fibrotic regions. As in the normal breast, fl-1 and a-6 immunoreactivity predominated in the processes of myoepithelial cells; faint fl-1 reactivity was again noted in the basolateral aspect of some luminal cells. Fibroadenomas and cystosarcoma phylloides Ductal structures were encompassed by a delicate line of laminin reactivity that faithfully followed the irregular contour of the myoepithelial cell layer; foci of reduplication were occasionally noted. Reactivity for Ten and cFn was noted in the stroma surrounding the ducts and extending unevenly into the distal stromal regions in a "raggedy" pattern similar to that observed in the fetal breast; the reactions were focally more intense in the cystosarcomas. Reactions for fl-1 and a-6 subunits of integrins were similar to the aforementioned. Ductal carcinomas Ducts containing in-situ carcinoma were encompassed by abnormal laminin reactivity ranging from rare multilayering to distinct attenuations to focal discontinuities. Ten and cFn were notably increased in amount and intensity, the former appeared as rings surrounding in-situ carcinomacontaining ducts while the latter was seen around ducts and in the stroma beyond them. More limited staining was also noted surrounding non-malignant but often hyperplastic ducts. Reactivity for fl-1 and a-6 was observed in the processes of residual, at times hyperplastic but presumably not malignant myoepithelial cells; weak fl-1 staining was seen in some of the intraductal malignant cells (Fig. le). IDC not otherwise specified displayed in most

413

instances scanty laminin reactivity limited to delicate lines, incompletely surrounding neoplastic clusters; some glandular structures were surrounded by laminin reactivity but most were not. Immunoreactions for Ten and cFn reactions were strong and extensive in the stroma associated with the carcinomas, distal stromal regions showed faint and limited staining. Weak a-6-specific cytoplasmic immunostaining was observed in variable majorities of tumor cells (Fig. lf); fl-1 staining involved fewer cells but was somewhat stronger and appeared enhanced at cell-cell and cell-matrix contact sites. Different reactions were noted in some variants of IDC, e.g. adenoid cystic carcinoma showed consistent laminin reactivity around neoplastic clusters and in the pseudocysts whereas mucinous carcinomas appeared entirely devoid of laminin staining but showed strong and rather extensive fl-1 staining (Fig. lg). Lobular carcinomas

In the in-situ areas, the distribution patterns of all molecules studied were similar to those noted in their ductal counterparts. ILC showed scanty but occasionally convincing laminin immunoreactivity surrounding clusters of tumor cells. Ten and cFn staining was strong and extensive in the tumors' stroma including some desmoplastic areas; the reactions were most intense in areas immediately adjacent to the tumors. Reactivity for a-6 was noted in 20-30% of tumor cells and predominated in delicate, branching cytoplasmic processes of single neoplastic cells; fl-1 staining showed a similar pattern and often involved a larger proportion of cells (Fig. lh).

Discussion

The spectrum of hyperplastic a n d / o r dysplastic breast lesions conventionally encompassed under the designation of FCD includes some particularly proliferative variants known for being at "high risk" for the subsequent development, or concomitant presence, of carcinoma (Page, 1986). Recent studies showed that the oncodevelopmental-type A-80 antigen richly expressed in breast carcinomas and weakly or not at all expressed in the normal

breast is indeed enhanced in FCD, and particularly in its proliferative forms (Gould et al., 1988; Koukoulis et al., 1991). Notably, FCD involves not merely epithelial proliferation but also significant stromal reorganization and remodelling as indicated by the occasionally conspicuous, desmoplastic (radial) "scars" found in certain FCD as well as in many carcinomas (McDivitt, 1990). In the context of these findings and given the known role of Ten in the developing breast, it was rather surprising that in adults Ten should only be found in the stroma of breast carcinomas (Mackie et al., 1987a). We can now say that while the accumulation of Ten is greatest in the stroma of breast carcinomas, it can also be detected through the spectrum of FCD, and is strongly enhanced in in-situ carcinomas (Howeedy et al., 1990). Furthermore, Ten's expression is not limited to pathologic hyperplasias but is noted in the classical example of physiologic hyperplasia, i.e. during pregnancy and lactation. Moreover, Ten is but weakly expressed in the normal resting adult and elderly breast. Therefore, the distribution of Ten in the breast is reminiscent of that of other ECM molecules in that it is not an "all or nothing" phenomenon (Howeedy et al., 1990). It is difficult to compare our findings on the distribution of EDAcFn in the breast with previous observations on the distribution of fibronectin, as many initial studies were carried out with antisera whose reactivity with regard to plasma vs cellular fibronectin was not defined (Christensen et al., 1988; Stampfer et al., 1981). As defined with the mAbs 100EB2 and 52DH1, respectively, the distribution patterns of Ten and cFn in the breast display parallels but also a number of differences. In fetal samples, EDAcFn did not form distinct periductal bands but was rather diffuse ensheathing most stromal cells. In the normal resting and gestational breast, cFn was noted in the periductal and interlobular matrix while being faint or absent in the intralobular stroma. In FCD, fibroadenomas and in in-situ carcinomas, immunoreactivity for EDAcFn staining was noted around ducts but also in the distal stroma while reactivity for Ten was most notable in the periductal stroma. In IDC and ILC, stromal reactions for cFn and Ten were similar except that, almost as a rule, they

414

were stronger and more extensive for the former while in the case of the latter, the distribution was more consistently restricted to the intratumoral and immediate peritumoral stroma. The distribution patterns of Ten and EDAcFn display a number of similarities, whereas a comparison between them and that of laminin reveals more complex correlations. If we select the adult resting breast as the standard normal, we can state that Ten and EDAcFn progressively increase in FCD, benign tumors and carcinomas while noting an initially apparently stable but subsequently increasingly defective, decreasing and finally virtually absent expression of laminin. In this context, our findings on laminin in the human breast and its neoplasms are similar to those of other investigators (Albrechtsen et al., 1981; Barsky et al., 1983; Christensen et al., 1989) while differing from findings in certain murine tumors (Pitelka et al., 1980). Our initial immunohistochemical observations with regard to the presence of the fl-1 and a-6 subunits of integrins in the normal, dysplastic and transformed human breast show some similarities with data obtained from in vitro studies (Ruoslahti and Giancotti, 1989; McDonald, 1989; Guirguis et al., 1987). Particularly pertinent are the cytokine-induced 67 kDa laminin and cFn receptor rich pseudopodia observed by Guirguis et al. (1987) in the MDA 435 human breast carcinoma cell line. Their description appears to parallel the conspicuous fl-1 and a-6 reactivity we noted in ILC as well as the variable distribution of the pertinent ligands. The apparent consistency of these events would suggest that the development of pseudopodia in these carcinoma cells, their interaction with cell adhesion molecules, and the differential distribution of ECM molecules in the adjacent vicinity may impact on cell locomotion in vivo. Therefore, given the potential functional significance of these morphologic findings, we propose for these structures the designation of kinetopodia. Furthermore, the different distribution of these molecules in IDC not otherwise specified vs their mucinous and adenoid cystic variants vs ILC would reflect a certain molecular diversity in these carcinomas which, in turn, may help explain well known, but nevertheless, poorly

understood differences in their local invasiveness and metastatic behaviour. Fibronectin receptor function has been considered to be involved in the metastatic process on the basis of an experimental melanoma model (Humphries et al., 1986). It should be noted that fl-1 integrin subunit can complex at least with six different a subunits, including a-6. However, our preliminary findings with mAbs to a-l, -2, -4 and -5 subunits show that they are not expressed in mammary carcinomas thus suggesting that the fl-1 immunoreactivity of tumor cells may correspond mainly to the a-3 subunlt.

Acknowledgements The original studies were supported by grants from the Otho S.A. Sprague Memorial Fund, the Finnish Academy of Sciences, the Finnish Cancer Research Fund and the Sigrid Jus61ius Foundation.

References Albrechtsen, R., Nielsen, M., Wewer, U., Engvall, E. and Ruoslahti, E.: Basement membrane changes in breast cancer detected by immunohistochemical staining for laminin. Cancer Res., 41, 5076-5081 (1981). Aufderheide, E., Chiquet-Ehrismann, R. and Ekblom, P.: Epithelial-mesenchymal interactions in the developing kidney lead to expression of tenascin in the mesenchyme. J. Cell Biol., 105, 599-608 (1987). Aufderheide, E. and Ekblom, P.: Tenascin during gut development: appearance in the mesenchyme, shift in molecular forms and dependence on epithelial-mesenchymal interactions. J. Cell Biol., 107, 2341-2349 (1988). Barsky, S.H., Siegal, G.P., Jarmotta, F. and Liotta, L.A.: Loss of basement membrane components by invasive tumors but not by their benign counterparts. Lab. Invest., 49, 140-147 (1983). Bourdon, M.A., Coleman, R.E., Blasberg, R.G., Grothius, D.R. and Bigner, D.D.: Monoclonal antibody localization in subcutaneous and intracranial human glioma xenografts: paired-label and imaging analysis. Anticancer Res., 4, 133140 (1984). Bourdon, M.A., Matthews, T.J., Pizzo, S.V. and Bigner, D.D.: Immunohistochemical and biochemical characterization of a glioma-associated extracellular matrix glycoprotein. Cell, 28, 183-195 (1985). Bourdon, M.A., Wikstrand, C.J., Furthmayer, H., Matthews,

415 T.J. and Bigner, D.D.: Human glioma-mesenchymal extracellular matrix antigen defined by monoclonal antibody. Cancer Res., 43, 2796-2805 (1983). Bronner-Fraser, M.: Distribution and function of tenascin during cranial neural crest development in the chick. J. Neurosci. Res., 21, 135-147 (1988). Chiquet-Ehrismann, R.: What distinguishes tenascin from fibronectin. FASEB J., 4, 2598-2604 (1990). Chiquet-Ehrismann, R., Mackie, E.J., Pearson, C.A. and Sakakura, T.: Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell, 47, 131-139 (1986). Christensen, L., Nielsen, M., Andersen, J. and Clemmensen, I.: Stromal fibronectin staining pattern and metastasizing ability of human breast carcinomas. Cancer Res., 48, 6227-6233 (1988). Christensen, I., Nielsen, M., Clemmensen, I. and Hage, E.: The immunohistochemical distribution of laminin and fibronectin in human female breast cancer. In Fenoglio-Preiser, C.M., Wilff, M., Rilke, F. (eds.), Progress in Surgical Pathology, Vol. 6, Field & Wood, Philadelphia, pp. 97-114 (1989). Crossin, K.L., Hoffman, S., Grnmet, M.S., Thiery, J.P. and Edelman, G.M.: Site-restricted expression of cytotactin during development of the chicken embryo. J. Cell Biol., 102, 1917-1930 (1986). Ekblom, P. and Aufderheide, E.: Stimulation of tenascin expression in mesenchyme by epitheial-mesenchymal interactions. Int. J. Dev. Biol., 33, 71-79 (1989). Gould, V.E., Shin, S.S., Manderino, G.L., Rittenhouse, H.G., Tomita, J.T. and Gooch, G.T.: Selective expression of a novel mucin-type glycoprotein in human tumors. Immunohistochemical demonstration with mAb A-80. Hum. Pathol. 19, 623-627 (1988). Gould, V.E., Koukoulis, G.K., Jansson, D.S., Nagle, R.B., Franke, W.W. and Moll, R.: Coexpression patterns of vimentin and glial filament protein with cytokeratins in the normal, hyperplasic and neoplastic breast. Am. J. Pathol., 137, 1143-1155 (1990). Grobstein, C.: Tissue interactions in the morphogenesis of mouse embryonic rudiments in vitro. In Rudnik, D. (ed.), Aspects of Synthesis and Order in Growth, Princeton University Press, Princeton, pp. 223-256 (1954). Grumet, M.S., Hoffman, S., Crossin, K.L. and Edelman, G.M.: An extracellular matrix protein of neural and non-neural tissues that mediates glia-nenron interaction. Proc. Natl. Acad. Sci. USA, 82, 8075-8079 (1985). Guirguis, R., Marguelis, I., Tarabolleti, G., Schiffman, E. and Liotta, L.A.: Cytokine-induced pseudopodial protrusion is coupled to tumor cell migration. Nature, 329, 261-263 (1987). Hay, E.: Extracellular matrix, cell skeletons, and embryonic development. Am. J. Med. Genet., 34, 14-29 (1989). Howeedy, A.A., Virtanen, I., Laitinen, L., Gould, N.S., Koukoulis, G.K. and Gould, V.E.: Differential distribution of tenascin in the normal, hyperplastic and neoplastic breast. Lab. Invest., in press (1990). Hsu, S.M., Raine, L. and Fanger, H.: Use of avidin-biotin-per-

oxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem., 29, 577-580 (1981). Humphries, M.J., Olden, K. and Yamada, K.M.: A synthetic peptide from fibronectin inhibits experimental metastasis of murine melanoma cells. Science, 233, 467-470 (1986). Inaguma, Y., Kusakabe, M., Mackie, E.J., Pearson, C.A., Chiquet-Ehrismann, R. and Sakakura, T.: Epithelial induction of stromal tenascin in the mouse mammary gland from embryogenesis to carcinogenesis. Dev. Biol., 128, 245-255 (1988). Koukoulis, G.K., Gould, V.E., Bhattacharyya, A., Gould, J.E., Howeedy, A.A. and Virtanen, I.: Tenascin in normal, reactive, hyperplastic and neoplastic tissues: biological and pathological implications. Hum. Pathol., in press (1991). Koukoulis, G.K., Jao, W., Shin, S.S., Gooch, G.T., Manderino, G.L., Rittenhouse, H.G., Tomita, J. and Gould, V.E.: Immunohistochemical evaluation of neoplastic and non-neoplastic breast diseases with mAb A-80. Pathol. Res. Pract., 186, 439-449 (1990). Kratochwil, K.: Organ specificity in mesenchymal induction demonstrated in the embryonic development of the mammary gland of the mouse. Dev. Biol., 20:46-71 (1969). Krnse, J., Keihauer, G., Faissner, A., Timpl, R. and Schachner M.: The JI glycoprotein-a novel nervous system cell adhesion molecule of the LS/HNK-I family. Nature, 316, 146-148 (1985). Liesi, P., Dahl, D. and Vaheri, A.: Laminin is produced by early rat astrocytes in primary culture. J. Cell Biol., 96, 920-924 (1983). Liotta, L.A.: Tumor invasion and metastases: role of the basement membrane. Am. J. Pathol., 117, 339-348 (1984). Mackie, E.J., Chiquet-Ehrismann, R., Pearson, C.A., Inaguma, Y., Taya, K., Kawarada, Y. and Sakakura, T.: Tenascin is a stromal marker for epithelial malignancy in the mammary gland. Proc. Natl. Acad. SCI. USA, 84, 4621-4625 (1987). Mackie, E.J., Halfter, W. and Liverani, D.: Induction of tenascin in healing wounds. J. Cell Biol., 107, 2757-2767 (1988a). Mackie, E.J., Thesleff, I. and Chiquet-Ehrismann, R.: Tenascin is associated with chondrogenic and osteogenic differentiation in vivo and promotes chondrogenesis in vitro. J. Cell Biol., 105, 2569-2579 (I987b). Mackie, E.J., Tucker, R.P., Halfter, W., Chiquet-Ehrismarm, R. and Epperlein, H.H.: The distribution of tenascin coincides with pathways of neural crest migration. Development, 102, 237-250 (1988b). Martinez-Hernandez, A.: The extracellular matrix and neoplasia. Lab. Invest., 58, 609-612 (1988). Martinez-Hernandez, A. and Amenta, P.S.: The basement membrane in pathology. Lab. Invest., 48, 656-677 (1983). McDivitt, R.: Breast, In Kissane, J.M. (ed.), Anderson's Pathology, CV Mosby, St. Louis, pp. 1726-1750 (1990). McDonald, J.A.: Receptors for extracellular matrix components. Am. J. Physiol., 257, 1331-1337 (1989). Page, D.L.: Cancer risk assessment in benign breast biopsies. Hum. Pathol., 17, 871-874 (1986). Pitelka, D.R., Hamamoto, S.T. and Taggart, B.N.: Basal lamina

416 and tissue recognition in malignant mammary tumors. Cancer Res., 40, 1600-1611 (1980). Riou, R.-S., Shi, D.-L., Chiquet, M. and Boucaut, J.-C: Expression of tenascin in response to neural induction in amphibian embryos. Development, 104, 511-524 (1988). Ruoslahti, E. and Giancotti, F.G.: Integrins and tumor cell dissemination. Cancer Ceils, 1, 119-126 (1989). Sakakura, T.: Epithelial-mesenchymalinteractions in mammary gland development and its perturbation in relation to tumorigenesis. In Rich, M.A., Hager, J.C., Furumanski, P. (eds.), Understanding Breast Cancer, Clinical and Laboratory Concepts, Marcel Dekker Inc., New York, pp. 261-284 (1983). Sonnenberg, A., Janssen, H., Hogervorst, F., Calafat, J. and Hilgers, J.: A complex of platelet glycoproteins Ic and IIa identified by a rat monoclonal antibody. J. Biol. Chem., 262, 10376-10383 (1987). Stampfer, M.R., Vlodavsky, I., Smith, H.S., Ford, R., Becker, F.F. and Riggs, J.: Fibronectin production by human mammary carcinoma cells. JNCI 67, 253-258 (1981).

Taylor, H.C., Lightner, V.A., Beyer, W.F., Jr., McCaslin, D., Briscoe, G. and Erickson, H.P.: Biochemical and structural studies of tenascin/hexabrachion proteins. J. Cell Biochem., 41, 71-90 (1989). Thesleff, I., Mackie, E.J, Vainio, S. and Chiquet-Ehrismann, R.: Changes in the distribution of tenascin during tooth development. Development, 101,289-296 (1987) Vartio, T., Laitinen, L., N~irv~lnen, O., Cutolo, M., Thornell, L.-E., Zardi, L. and Virtanen, I.: Differential expression of the ED sequence-containing form of cellular fibronectin in embryonic and adult human tissues. J. Cell Sci., 88, 419-430 (1987). Wewer, U., Albrechtsen, R., Manthorpe, M., Varon, S., Engvall, E. and Ruoslahti, E.: Human laminin isolated in a nearly intact, biologically active form from placenta by limited proteolysis. J. Biol. Chem., 258, 12654-12660 (1983). Yl~inne, J. and Virtanen, I.: The M r 140,000 fibronectin receptor complex in normal and virus-transformed fibroblasts and in fibrosarcoma cells: identical localization and function. Int. J. Cancer, 43, 1126-1136 (1989).