APMIS 98: 6 15-623, 1990

Fibronectin: A discrimination marker between small invasive carcinomas and benign proliferative lesions of the breast LISE CHRISTENSEN The Department of Pathology, Rigshospitalet, University of Copenhagen, Denmark

Christensen. L. Fibronectin: A discrimination marker between small invasive carcinomas and benign proliferative lesions of the breast. APMIS 98: 6 15-623, 1990. The connective tissue glycoprotein, fibronectin (FN) is known to be increased in the stromal desmoplastic response associated with invasive breast carcinoma (IBC). In a previous study of 77 IBCs we found that all tumours. irrespective of desmoplasia, displayed an intense stromal immunoreactivity for FN, although those with a high metastatic potential tended to lack this in foci along their infiltrating border (Christensen et al. 1988). In order to test the diagnostic value of this observation, FN immunoreactivity was estimated semiquantitatively in I3 1 small primary IBCs, 23 in situ carcinomas, 20 borderline lesions, 142 benign proliferative breast lesions and 35 samples of normal breast tissue. All IBCs, including those without apparent desmoplasia, were accompanied by an intense stromal staining for FN forming a diffuse or pericellular pattern around the tumour cells (FN positive reaction). This was not seen in the benign lesions (FN negative reaction) with the exception of an FN positive, linear staining around some cysts and ectatic ducts and a diffuse staining of the connective tissue core of some intraductal papillomas. Intraductal carcinomas, multiple papillomas and radial scars could display a substantial staining for FN, sometimes enclosing small islands or tubules of epithelial cells. However, the fact that all IBCs of this study produced an overall reliable and consistent FN positive stromal reaction, which was only interrupted in minute foci along the infiltrating border in a few of the largest tumours ( 1 I / 13 I), makes immunohistochemical staining for FN an excellent adjunct to morphology in tracing small, primary IBCs. Key words: Fibronectin: differentiation marker: small invasive breast carcinomas: benign breast lesions. L. Christensen, Department of Pathology, Rigshospitalet, Fred. V's vej 1 I , 2100 Kerbenhavn 0, Denmark.

Fibronectin (FN) is a structural, non-collagenous glycoprotein of connective tissues, basal laminas and blood, which has been identified intra- and extracellularly as well as in the growth media of various epithelial and mesenchymal cells in vitro (for review, please consult reference no. I). It appears to participate in tissue formation and predominates quantitatively in early granulation tissue and during embryonic development ( 12, 14, 31). Normal breast tissue contains only a few,

Received July 18, 1989. Accepted January 15, 1990.

scattered fibres, which stain immunologically for FN (2, 7, 10,22, 27). Their numbers are, however, significantly increased during inflammatory and proliferative states and form a dense network in the desmoplastic stromal response, which is characteristic of many invasive breast carcinomas (IBC) (7, 10, 17, 19, 23, 28). In a previous study of 77 primary IBCs we discovered that FN was increased in all tumours, including those without a morphologically recognizable desmoplastic reaction, although focal dropouts occurred along the infiltrating border in tumours with a high metastatic potential (6) and large size (own unpublished observation). Antici615

CHRISTENSEN

pating that small tumours retain a high degree of clonal homogeneity with a minimum loss of stromal FN expression, the present study was designed to investigate whether intense stromal FN immunoreactivity is a constant characteristic of small IBCs to the extent that it can be of value in differentiating between benign and malignant epithelial breast tissue. MATERIALS AND METHODS Tissues Formalin-fixed, parafin-embedded tissues from 13 I small primary invasive breast carcinomas, 23 in situ carcinomas, 20 borderline lesions, 142 benign proliferative breast lesions and 35 normal breast samples were obtained from files of the Departments of Pathology, Rigshospitalet and Frederiksberg Hospital, University of Copenhagen, The University Hospital of Odense, Svendborg Hospital and Randers Hospital, Denmark. The carcinomas were selected according to size, none of them exceeding 10 mm. Normal breast tissue originated from breasts operated upon because of hypertrophia or ptosis. The other benign lesions had been removed due to clinical and/or radiological suspicion of carcinoma. All tissues had been fixed in 10% neutral buffered formalin, routinely dehydrated and embedded in parafin at 58 "C. Five pm sections were cut and stained routinely with H&E, van Gieson/Alcian blue, Gordon and Sweets reticulin stain and orcein. Additional neighbouring sections were stained immunohistochemically for FN and actin, respectively. Imm unoperoxiduse Stuining .for FN A standardized indirect peroxidase-antiperoxidase (PAP) staining technique was used (Table 1) (5). Prior to immunostaining deparaffnized sections were treated with 1% testicular hyaluronidase (Hyalase) (LEO, Sweden) in McIlvain buffer, pH 7 for 30 rnin at 37 "C (1 5) followed by 0.1% protease (Sigma, St. Louis, Mo, VII 5255) in Tris buffered saline at room temperature for 5 rnin (5). The primary specific antibody was a polyclonal IgG immunoglobulin raised in rabbits against human fibronectin (Dakopatts Ltd., code no. A 245, lot no. 091B) at a protein concentration of 50 pg/L The secondary antibody was swine anti-rabbit immunoglobulin (SWAR) (Dakopatts Ltd., code no. Z 195) used at a concentration of 1:20, and the tertiary antibody was rabbit peroxidase-antiperoxidaseIgG (PAP) (Dakopatts Ltd., code no. Z 1 13. lot no. 053). As negative controls were used 1 ) anti-FN absorbed with purified FN antigen (kindly donated by Dr I. Clemmensen, Department of Clinical Microbiology, Statens Serum Institut at Rigshospitalet, University of Copenhagen, Denmark), and 2) replacement of the primary specific antibody with the IgG-fraction of serum from non-immunized rabbits (Dakopatts Ltd., code no. X 903, lot no. 038 B).

616

Imm unoperoxiduse Stuining.for Act in Neighbouring sections to those stained for FN were stained for actin by the indirect triple-layer immunoperoxidase staining technique using a mouse monoclonal antibody (Amersham Bucks, code no. N 350, lot no. B 10, dilution 1:800). The staining procedure has been described previously (5). In short it includes the following steps: I ) After deparafinization and hydration, sections are proteolytically digested with 0. I % pronase in TBS for 2 min, 2) endogenous peroxidase activity is blocked with 1% HzOz in methanol for 20 min, 3) sections are preincubated with 1% human albumin in TBS for 5 rnin followed by incubation for 30 min with the primary, specific antibody, 4)incubation with peroxidase-conjugated rabbit anti-mouse immunoglobulin (Dakopatts Ltd., code no. P 260, dilution 1:50) in TBS/albumin for 30 min, 5) incubation with peroxidaseconjugated swine anti-rabbit immunoglobulin (Dakopatts Ltd., code no. P 217, dilution 1:20 in TBS/albumin). Washing between incubations, development and mounting of the slides were carried out as for FN immunostaining.

TABLE 1. Immunoperoxiduse staining protocol for jbronectin on formalin-Jixed tissue 1. Deparafinize and hydrate to water 2. Proteolytic pretreatment with 0. I % pronase in TBS 3. Hyalase, 1 promille, in McIlvain buffer at 37 "C 4. Block endogenous peroxidase activity by 1% HzOzin methanol 5. Preincubate in 1Yo human albumin in TBS 6. Allow excess of preincubating medium to run off sections 7. Incubate with polyclonal rabbit antibody to human FN 8. Apply swine-anti-rabbit immunoglobulin diluted 1:20 in 1% human albumin in TBS 9. Apply peroxidase-antiperoxidase complex (PAP) diluted I :40 in 1% human albumin in TBS LO. Develop in 3-amino-9-ethylcarbazole solution I 1. Counterstain in Mayer's hemalun 12. Rinse in distilled water and mount in Glycergel (Dakopatts Ltd.)

5 min 30 min 20 min 5 min

30 min 30 min 25 rnin 10 min 5 min

Between the points 2 to 5 and 7 to 10 sections were washed 2 x 5 rnin in TBS/Triton X- 100 (190).Between points 10 to 12 sections were rinsed in running tap water. If not otherwise indicated, the reactions were carried out at room temperature. TBS: Tris buffered saline (50 mlO.l M Tris buffer, pH 7.6 in 1 10.9% NaCI.

FIBRONECTIN IN BREAST TUMOURS

Classification and Evaluation Morphology: The original H&E sections were reviewed and reclassified by the author prior to immunostaining. Benign and borderline breast lesions as well as in situ carcinomas were diagnosed according to the criteria of Azzopardi (3), Haagensen ( 13)and Line11( 2 0 ) , and the IBCs were histologically classified and subclassified by following the guidelines of the World Health Organization (WHO) ( 2 6 )and Bloom &Richardson (4), respectively. FN positive stroma staining was recorded if the stroma showed a tine or coarse network of FN positive fibres (diffuse pattern), or if an intense, linear, basement membrane-like staining for FN was present around individual cells or groups of cells (pericellular pattern). FN negative stromal staining was defined as the presence of a few, scattered, FN positive strands within the connective tissue. Interlobular normal breast connective tissue was used as reference.

TABLE 2. Histological type and FN staining pattern of 142 benign proliferative breast lesions and 43 borderline lesions Histological type

Benign lesions cysts duct ectasia adenosis sclerosing adenosis adenosis tumor fibroadenoma lactating adenoma tubular adenoma epit heliosis solitary papilloma Total

RESULTS

The FN positive stromal staining was intense and specific and provided a sharp contrast to the surrounding FN negative stroma of normal-looking breast tissue (Figs. 2-10). Small blood vessels, some of which were easily identified by a fine, linear, subendothelial staining, served as built-in positive controls for the efficiency of the staining method. Luminal contents and serum in blood vessels showed a weak, amorphous staining. Extravasated blood was homogenously stained for FN. Also foci of early granulation tissue, observed in connection with ruptured ducts, showed an amorphous FN positive reaction, whereas chronic inflammatory changes were unaccompanied by such a pattern. Fibroblasts showed a moderate and myofibroblasts and intense cytoplasmic immunoreactivity for FN. Also some of the tumour cells were weakly stained, but the majority appeared FN negative. Actin positive stromal cells (fibroblasts and myofibroblasts), were present in increased numbers whenever desmoplasia was apparent. Smooth muscle cells and myoepithelial cells were actin positive as well, whereas all epithelial cells, including the tumour cells, appeared actin negative by the staining procedure used.

Normal Breast Tissue Normal breast tissue and normal-looking breast tissue adjacent to the IBCs showed a consistent, FN negative staining reaction (Fig. 1). Slight nuances were observed between intra- and interlobular

FN positive staining 6 4

0 0 I* 0 0 0 0 (2) 12

FN negative staining 8 11 15 32 4

18 7 3 25 7 130

*Morphological suspicion of transition into infiltrating duct carcinoma.

Borderline lesions radial scar multiple papilloma In situ carcinomas ductal in situ carcinoma a) comedotype b) cribriform type lobular in situ carcinoma Total

6 7

7 0

6 2 0

4 3

22

21

8

connective tissues, where the former tended to display a few more FN immunoreactive fibres than the latter, but the difference was marginal (Fig. 1).

Benign Proliferative Breast Lesions (Table 2) Several benign changes were often present within the same tissue, but only the dominating type of lesion was registered. Cysts and dilated ducts were in some cases surrounded by a circle of FN positive fibres (Fig. 2). These were running parallel to the epithelial lining, suggesting mechanical compression of normal connective tissue fibres rather than increased production of FN. Also some of the solitary duct papillomas displayed a fine, FN positive lattice corresponding to their connective tissue core - perhaps due to a similar compresion mechanism (Fig. 3). Fibroadenomas, tubular and lactating adenomas as well as foci of adenosis were consistently FN negative. However, one adenosis tumour with morphological suspicion of transition into a grade 1 duct carcinoma displayed 617

CHRISTENSEN

Figs. I - 11: Formalin-fixed, paraffin-embedded breast tissue stained for fibronectin (FN) by the indirect peroxidase-antiperoxidase technique (PAP) and counterstained with Mayer’s hemalun. Fig. 1. FN negative normal resting breast tissue showing a few more immunoreactive fibres intralobularly than in the surrounding interlobular connective tissue (PAP x 160). Fig. 2. FN positive reaction surrounding a benign cyst. The immunoreactive fibres run parallel to the epithelial lining suggesting increased staining due to mechanical compression of normal connective tissue (PAP x 160). Fig. 3. Solitary duct papilloma displaying a fine, diffuse FN positive reaction corresponding to the connective tissue core. Apart from subendothelial staining of a major vessel (top left) the surrounding stroma is negative (PAP x 160). Fig. 4. Peripheral part of an adenosis tumor with morphological suspicion of transition into a well-differentiated duct carcinoma. A diffuse, fine, distinctly FN positive staining pattern is apparent (PAP x 160). Fig. 5. Multiple papillomas surrounded by an FN positive stroma (a) containing a few, small epithelial islands with cellular features like those of the papilloma (b), (PAP x 86 (a) and x 210 (b)).

618

FlBRONECTlN IN BREAST TUMOURS

Fig. 6. Radial scar with an FN positive staining corresponding to the central area containing small, compressed, tubular structures. Nodules of (orcein positive) elastin are negative (arrow) (PAP x 160). Fig. 7. Intraductal carcinoma with cancerization of lobules embedded in an FN positive stroma. The terminal acini vary in size, and small. solid epithelial islands seem to bud off from these, suggesting early invasion (PAP x 160). Fig. 8. Well-differentiated invasive duct carcinoma displaying an FN positive staining reaction of the pericellular type (PAP x 160). Fig. 9. Grade 2 duct carcinoma showing a diffuse, coarse stromal staining for FN (PAP x 160). Fig. IO. Atypical medullary carcinoma displaying a pencellular staining pattern for FN,surrounding small groups or individual cells within the solid sheets of the tumour (PAP x 200). Fig. I I . Vascular tumour embolisms surrounded by an irregular cuff of FN positive fibres. The stroma, which contains numerous fibroblastic cells and chronic inflammatory cells, is unreactive (top) (PAP x 200).

619

CHRISTENSEN

stromal FN positivity corresponding to this area (Fig. 4).

Borderline Breast Lesions (Table 2) Multiple papillomas were, like cysts, sometimes surrounded by a ring of FN positive fibres, and like solitary papillomas their connective tissue core was stained as well. Frequently, however, larger foci of stromal FN positivity occurred, which seemed to reflect increased synthesis of FN rather than mere compression of normal connective tissue fibres (Fig. 5a+b). A few tubular or solid epithelial islands, which in cellular characteristics did not deviate from those of the papillomas, and which were inconsistently surrounded by actin positive myoepithelial cells, were generally present in this coarse, FN positive stroma (Fig. 5b). Radial scars displayed variations in FN staining pattern as well. Typical lesions with a small, central, elastic core surrounded by radiating, dilated ducts were sometimes intensely stained for FN, especially when tubular structures of the type described in connection with multiple papillomas were present (Fig. 6).

In Situ Carcinomas (Table 2) Fifteen tumours were classified morphologically as pure ductal in situ carcinomas and 8 as pure lobular in situ carcinomas. All of the lobular and most of the ductal in situ elements were devoid of stromal FN positivity, but sometimes the latter were surrounded by an irregular ring of FN posi-

tivity, much like the one observed around cysts and dilated ducts. This was especially seen in cases of lobular cancerization (Fig. 7). Central necroses of so-called comedoelements from ductal in situ carcinomas were sometimes weakly stained.

Invasive Carcinomas (Table 3) All IBCs, irrespective of type and size, showed a convincing, FN positive stromal staining reaction (Figs. 8-10). The FN positivity was in general equally distributed throughout the tumour tissue, presenting only one type of staining pattern (diffuse or pericellular) in 6 1 of the smallest tumours (Figs. 8- 10)and a combination of staining patterns in the other 70 (Table 3). Small foci without FN reactivity along the infiltrating border were present in five grade 2 duct carcinomas, one grade 3 duct carcinoma and in five invasive lobular carcinomas, all among the largest tumours. Four tumours were associated with small vascular tumour embolisms. They were surrounded by a thick, FN positive rim corresponding to the vessel wall and its periphery. The intervening stroma, which in 1 case contained chronic inflammatory cells and numerous fibroblasts, was unreactive (Fig. 11). No specific type of staining pattern was associated with the ductal carcinomas (Table 3). This was also true for the tubular carcinomas, whereas the 3 atypical medullary carcinomas tended to stain around individual tumour cells in spite of a solid growth pattern in sheets (Fig. 10). The majority of invasive lobular carcinomas displayed

TABLE 3. Histological type and type of FN staining pattern of 131 IBCs Histological type

Type of FN

Positive staining

Pattern

Diffuse

Pericellular

Mixed

duct carcinoma grade 1 duct carcinoma grade 2 duct carcinoma grade 3 tubular carcinoma invasive lobular carcinoma invasive lobular/duct carcinoma papillary carcinoma adenoid cystic carcinoma signet ring cell carcinoma colloid carcinoma atypical medullary carcinoma

13

5

2 0

1

Total

27

620

0

19 13 2 5 16 11 1 0 3 0

0 34

70

FIBRONECTIN IN BREAST TUMOURS

the pericellular type of staining pattern surrounding individual cells as well as rows of tumour cells forming the so-called “single filing” pattern. DISCUSSION Various immunohistochemical studies have demonstrated an increased presence of tissuebound FN in IBC (2, 7, 10, 23, 28), but it has not been determined if this is present in all IBCs to the extent that stromal immunoreactivity for FN can be of diagnostic value. In a previous study of 77 clinically manifest, primary IBCs we noticed that all tumours, including those without apparent desmoplasia, displayed a significantly increased stromal immunoreactivity for FN in comparison to the surrounding normal-looking breast tissue (6). However, several of the tumours, both with and without desmoplasia, contained FN negative foci along their infiltrating border, which seemed to be significantly associated with metastatic disease (6). The purpose of the present study was to test the usage of FN in the identification of incipient, primary IBCs as opposed to benign, proliferative lesions. Therefore, only tumours under 10 mm were considered, resulting in a low incidence of FN negative foci occurring in only 1 1 out of 131 tumours - all among the largest ones. In another recent study of FN immunoreactivity in IBCs, stromal FN was found to be minimal or absent in 12 out of 18 invasive lobular carcinomas (7). This is in discordance with the findings of the present study and may be due to the fact that the FN positive staining pattern of the pericellular type seen in most invasive lobular carcinomas is only reliably expressed in tissue sections that have been subjected to hyaluronidase digestion prior to immunostaining (1 5). In benign, non-embryonic tissues, stromal FN is increased during the early stages of granulation tissue and in tissue repair (12, 14, 22). Yet, even such tissues seem to express quantitative differences in benign and malignant lesions, as a 30times increase in FN production has been measured in scar tissue adjacent to carcinoma as compared to scar tissue not associated with carcinoma ( 16). In the present study we were not able to confirm this observation. Sporadic, morphologically characteristic, reparative lesions of FN positive early granulation tissue with granulocytes, mainly seen in connection with ruptured cysts,

could not be distinguished from IBC by immunostaining for FN. The same was true for cysts, dilated ducts and solitary papillomas. In these cases morphology was essential for the diagnosis (Figs. 2 and 3). The linear FN positivity surrounding cysts and dilated ducts has been observed by others (23). It seems to reflect compressionof normal connective tissue rather than increased synthesis or incorporation of FN, as judged by the parallel organization of the fibres following the circumference of the ducts. Unfortunately, a similar staining was observed around elements of ductal carcinoma in situ, ruling out FN as a discrimination marker between such lesions and benign cysts. The FN positivity observed in connection with in situ carcinomas may also, at least partly, stem from an increased production of FN. An ultrastructural study has shown that gaps in the basement membrane surrounding the elements are associated with proliferation of fibroblastic cells (fibroblasts and myofibroblasts) (29), the main cells of desmoplasia (19). Such cells have been found to produce increased amounts of FN mRNA in response to transforming growth factor-p, a potent desmoplastic agent (30), which has been found to be produced by the estrogen-dependent breast cancer cell line, MCF-7, at an increased rate under the influence of anti-estrogens (1 8). That FN production is hormonally regulated and not just a reflection of the fibroblastic cells present in the tissue was in the present study illustrated by an FN negative fibroblast-rich stroma found between vascular tumour emboli (Fig. 11). Possibly this represented granulation tissue of advanced age, which is known to express little FN, as opposed to the early stages, which contain large quantities of the protein (14). However, the intense staining of the vessel wall and its immediate surroundings matches recent findings, which have shown that besides producing desmoplastic factors such as transforming growth factor-p ( 18), breast cancer cells produce proteolytic enzymes capable of degrading the extracellular matrix of IBC and influencing disease-free survival (8). Plasma FN may contribute to the increased, penvascular FN positive reaction (24). The more widespread FN positivity observed around multiple papillomas (Fig. 5 ) and intraductal carcinomas with cancerization of lobuli (Fig. 7) were unlikely to represent stromal compression alone. The occurrence of small epithelialislandsor tubules within this intensely stained, desmoplas62 1

CHRISTENSEN

tic-looking stroma gave the impression of an increased deposition or de novo production of FN (Figs. 5b and 7). Whether early reparative changes or neoplastic invasion are responsible remains to be determined. Radial scars, some of which displayed a similar FN positive reaction pattern centrally (Fig. 6), have been suspected of representing incipient breast carcinomas (9, 20). Technically, the staining procedure for detection of stromal FN as described in Table I is reliable and reproducible (3,and variations in processing conditions have little influence on the staining results. Only autopsy material and tumours with large necroses may cause an unwanted background staining which masks the reaction pattern (6). In such cases, the risk of detaching tissue sections from the glass slides during proteolytic pretreatment is increased as well (6). The contractile protein actin, which in breast tissue can be demonstrated immunohistochemically within smooth muscle cells, myofibroblasts, myoepithelial cells and some malignant tumour cells (2 I), represented in this study another indicator of desmoplasia. However, although the tumour cells were left unreactive by the staining procedure used, the myoepithelial cells were stained as well, rendering the protein inefficient when discriminating between benign proliferative lesions such as sclerosing adenosis and tubular carcinoma. FN immunoreactivity, on the other hand, was consistently increased in the stroma of IBC as opposed to most benign lesions and appeared to provide a valuable adjunct to morphology in tracing IBC. The diagnostic significance of this observation remains to be established in a clinically controlled study where a panel of experts in breast pathology review the sections blindly. Publication costs were supported by a grant from the Danish Medical Research Council and Dakopatts Ltd., Denmark. Professor, dr. med. J. Andersen, Department of pathology, University Hospital of Odense, Chief pathologist, dr. med H . Kjcer, Department of pathology, Svendborg Centralsygehus,and Senior Pathologist Bernt Nielsen , Department of Pathology, Randers Centralsygehus, are thanked for contributing some of the borderline lesions.

REFERENCES 1. Akiyama, S. K. & Yamada, K.: Fibronectin in disease. In: Wagner, B., Fleischmajer, R. & Kaufman, N. (Eds.): Connective tissue diseases. Balti-

622

more M. D.: Williams and Wilkins 1985, pp. 55-96. 2. Alon, Y.,Horowitz, A . T., Biran, S., Weiss, D. W. & Doljanski, F.: Immunofluorescent characterization of fibronectin, laminin and keratin in normal and neoplastic human mammary epithelial cells in culture and in breast tissue sections. Int. J. Tissue React. 8: 410-10, 1986. 3. Azzopardi, J. G.: Problems in Breast Pathology. Philadelphia: H. B. Saunders, 1979. 4. Bloom, H . J. G. & Richardson, W. W.: Histological grading and prognosis in breast cancer. A study of 1409 cases of which 359 have been followed for 15 years. Br. J. Cancer 11: 359-377, 1957. 5. Christensen, L. & Strange, L.: Universal immunoperoxidase staining protocol to optimize the use of polyclonal and monoclonal antibodies. J Histotechnol. 10: 11-15, 1987. 6. Christensen, L., Nielsen, M., Andersen, J. & Clemmensen, I.: Stromal fibronectin staining pattern and metastasizing ability of human breast carcinoma. Cancer Res. 48: 6227-6233, 1988. 7. D’Ardenne, A. J. & Barnard, N. J.: Paucity of fibronectin in lobular carcinoma of the breast. J. Pathol. 1-57; 219-224, 1989. 8. Duffy, M . J., O’Grady, P., Devaney, D., O’Siorain, L., Fennelly, J. & Kijnen, H . J.: Urokinase-plasminogen activator, a marker for aggressive breast carcinomas. Preliminary report. Cancer 62: 5-3 1-533, 1988. 9. Fisher, E. R., Palekar, A. S., Sass, R. & Fisher, B.: Scar cancers: Pathologic findings from the national Surgical Adjuvant Breast Project (Protocol No. 4): IV. Breast Cancer Res. Treat 3: 39-59, 1983. 10. Friedman, R., Gelfand, T., Weiss, D. W . & Doljanski, F.: Patterns of fibronectin deposition in normal and neoplastic fibroblasts and mammary tissue. Int. J. Tissue Reac. 6: 291-301, 1984. 1 1. Gallagher, H. S. & Martin, J. E.: Early phases in the development of breast cancer. Cancer 24: 11701178, 1969. 12. Grinnell, F., Billingham, R. E. &Burgess, L.: Distribution of fibronectin during wound healing in vivo. J. Invest. Dermatol. 76: 181-189, 1981. 13. Haagensen, C. D.: Diseases of the breast, 3rd ed., Philadelphia, London, Toronto: W. B. Saunders, 1986. 14. Hdund, B., Clemmensen, I., Junker, P. & Lyon, H.: Fibronectin in experimental granulation tissue. Acta path. microbiol. immunol. scand. Sect. A, 90: 159165, 1982. 15. Hdund, B. & Clemmensen, I.: The value of hyaluronidase treatment of different tissues before demonstration of fibronectin by the indirect immunoperoxidase technique. Histochemistry 76: 517-525, 1982. 16. Igisu, K.: Increased production of fibronectin by fibroblasts from precancerous scars. Brit. J. Dermatol. 116: 79-86, 1987. 17. Fao, R. T., Hall, J., Engel, L. & Starn, R.: The matrix of human breast cancer is mitogenic for fibroblasts.

FIBRONECTIN IN BREAST TUMOURS

Am. J. Parthol. 115: 1099-1 16, 1984. 18. Knabbe, C., Lippman, M . E. & Wakefield, L . M., Flanders, K. C.. Kasid, A.. Derynck, R. & Dickson, R. B.: Evidence that transforming growth factor-p is a hormonally regulated negative growth factor in human breast cancer cells. Cell 48: 417-428, 1987. 19. Lagace, R., Grimaud, J.-A., Schwurch, W. & Seemayer, T. A.: Myofibroblastic stromal reaction in carcinoma of the breast: Variations in collagenous matrix and structural glycoproteins. Virchows Arch (Pathol. Anat) 408: 49-59, 1985. 20. Line4 F., Ljunberg, 0. & Anderson, I.: Breast carcinoma. Aspects of early stages, progression and related problems. Acta pathol. microbiol. scand. Sect. A 1980, suppl. no. 272. 21. Mukai, K.. Schoilmeyer, J. V. & Rosai, J.: Immunohistochemical localization of actin. Am. J. Surg. Pathol. 5: 91-97, 1981. 22. Nagata, H., Ueki, H . & Moriguichi, T.:Fibronectin. Localization in normal human skin, granulation tissue, hypertrophic scar, mature scar, progressive systemic sclerotic skin and other fibrosing dermatoses. Arch. Dermatol. 121: 29-33, 1985. 23. Natali, P. G., Giacomini, P., Bigotti, G.,Nicotra, M. R., Bellocci, M. & De Martino, C.: Heterogeneous distribution of actin, myosin, fibronectin and basement membrane antigens in primary and metastatic human breast cancer. Virchows Arch. (Pathol Anat)

405: 69-83, 1984. 24. Oh, E., Pierschbacher, M. & Rouslahti, E.: Deposition of plasma fibronectin in tissues. Proc. Natl. Acad. Sci 787: 3218-3221, 1981. 25. Ozzello, L.:The behaviour of basement membranes in intraductal carcinoma of the breast. Am. J. Pathol. Xf35:887-899, 1959. 26. Sobin, L . H.: Histological typing of breast tumors, ed. 2. Geneva: WHO, 198 1 . 27. Stenman, S. & Vaheri, A.: Distribution of a major connective tissue protein, fibronectin, in normal human tissues. J. Exp. Med. 147: 1054-1064, 1978. 28. Stenman, S. & Vaheri, A.: Fibronectin in human solid tumors. Int J. Cancer 27: 427-435, 1981. 29. Tamini, S. 0.& Ahmed, A.:Stromal changes in early invasive and non-invasive breast carcinoma: An ultrastructural study. J. Pathol 150: 43-49, 1986. 30. Varga,J., Rosenbloom, J. & Jimenez, S. A,: Transforming growth factor-p (TGF-P) causes a persistent increase in steady-state amounts of type I and type I11 collagen and fibronectin mRNAs in normal human dermal fibroblasts. Biochem. J. 247: 597604, 1987. 3 1. Wartiovaara,P. & Vaheri,A.: Fibronectin and early mammalian embryogenesis. In: Johnson, M. H . (Ed.): Development in mammals. Elsevier: North Holland biomedical press, 1980, pp. 233-265.

623

Fibronectin: a discrimination marker between small invasive carcinomas and benign proliferative lesions of the breast.

The connective tissue glycoprotein, fibronectin (FN) is known to be increased in the stromal desmoplastic response associated with invasive breast car...
5MB Sizes 0 Downloads 0 Views