Tumor Cells of Malignant Fibrous - Europe PMC

American Jounal of Pathology, Vol. 139, No. 5, November 1991 Copright © American Association of Pathologists

Tumor Cells of Malignant Fibrous Histiocytomas Express mRNA for Laminin Ylermi Soini and Helena Autio-Harmainen From the Department of Pathology, University of Oulu, Oulu, Finland

In situ hybridization was used on routinely processed paraffin-embedded tissue sections to study the synthesis of the basement membrane (BM) proteins laminin and type IV collagen in 14 cases of malignant fibrous histiocytoma (MFH). Complementary RNA probes coding for the pro-ca (IV) chain of human type lV collagen and the BI chain of human laminin were used to detect the respective mRNAs. The results were correlated with the immunohistochemical reactivity of tumor cells to specific antibodies against the P1 fragment of laminin and the 7S domain of type IV collagen. Signals for the presence of laminin mRNA in atypical neoplastic tumor cells could be detected in 11 MFHs. None of the tumors could be shown to contain signals for type IV collagen mRNA in their cells, although such signals were detected in the endothelial cells of tumor capillaries. In the corresponding immunohistochemical stainings, nine MFHs showed intracytoplasmic staining of tumor cells for laminin and one tumor showed weak staining for type IV collagen in the neoplastic cells. The results show that the laminin immunoreactivity found in MFHs is due to synthesis in the tumor cells and not to endogenous uptake of this protein. Synthesis of laminin in the majority of MFHs is in accordance with the notion that these tumors originate from primitive mesenchymal cells in soft tissues. (Am J Pathol 1991, 139:1061-1068)

Malignant fibrous histiocytoma (MFH), the most common soft tissue tumor of adult life,1 is thought to originate from histiocytes,2 fibroblasts3'4 or primitive mesenchymal cells.5 Immunohistochemically the tumor cells have been shown to be positive for vimentin.6'7 Reactivity for other intermediate filament proteins such as desmin, cytokeratin, and neurofilament also has been described.6'7 Because most recent reports have found no immunoreactivity to monoclonal histiomonocytic markers in MFH tumor cells,668 their histiocytic origin has been questioned.

The issue is complicated, however, by reports describing positive reactions to histiomonocytic markers in a restricted number of MFHs.911' Laminin and type IV collagen are the basic constituent proteins of basement membranes (BMs),12 and are generally synthesized by cells invested with a BM.12 Extracellular reactivity for these proteins in soft tissues has been shown to be associated with endothelial cells, smooth and striated muscle cells, lipocytes, and Schwann cells.12 Consequently vascular, myogenic, Iipogenic, and neurogenic soft tissue tumors have been reported to be positive.13'14 Some earlier immunohistochemical investigations have shown MFHs to be negative for laminin and type IV collagen,13'14 but our previous investigation with material from 16 MFHs showed considerable intracytoplasmic laminin positivity in the tumor cells of most MFHs, and two of the tumors also expressed intracytoplasmic type IV collagen.15 It was proposed that this could be a reflection of the mesenchymal nonhistiocytic nature of these tumors.15 Exogenous uptake of these proteins by the tumor cells could not be excluded, however.15 In this investigation we analyzed 14 MFHs both immunohistochemically and by in situ hybridization techniques to establish if synthesis of these BM proteins could be detected at the mRNA level and how it correlated with the immunohistochemical results.

Materials and Methods Fourteen MFHs were collected from the files of the Departments of Pathology, Universities of Oulu and Helsinki, between 1981 and 1989. The clinical data on the cases are presented in Table 1. All the tumors had been fixed in 10% buffered formalin and embedded in paraffin. The material consisted of 12 pleomorphic-storiform, one giant cell, and one inflammatory subtype of MFH. The diagnosis of all tumors was based on light microscopy according to the criteria of Enzinger and Weiss.1 Most tumors also were studied by electron microscopy. Accepted for publication June 25, 1991. Address reprint requests to Dr. Ylermi Soini, MD, Department of Pathology, University of Oulu, Kajaanintie 52 D, SF-90220 Oulu, Finland.

1061

1062

Soini and Autio-Harmainen

AJP November 1991, Vol. 139, No. 5

Table 1. Clinical Data of the Cases Studied

Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Age 63 68 77 60 62 79 57 65 81 50 77 47 67 70

Sex

Location

Subtype

Male Female Male Female Female Male Male Male Male Female Female Male Female Female

Chest wall Upper extremity

GC PS PS PS PS PS PS PS PS IS

Thigh Clavicular region Axilla Thigh Calf Upper extremity Calf

Retroperitoneum Upper extremity

Calf Lung Upper extremity GC = giant-cell subtype; PS = pleomorphic-storiform subtype; IS - inflammatory subtype.

Treatment of Sections for In Situ Hybridization For in situ hybridization, paraffin sections 4 pu in thickness, cut onto a flotation bath containing 1% Elmer's glue.1 6-18 The sections then were transferred to microscope slides that had been treated with poly-L-lysine (50-100 ,ug/ml). Before hybridization, the sections were baked at 370C overnight and at 600C for 2 hours, after which the slides were directly transferred to xylene (2 x 10 minutes), 100% ethanol (1 x 15 minutes), 96% ethanol (1 x 5 minutes), and 70% ethanol (1 x 5 minutes) and washed in PBS for 3 minutes. The sections were allowed to dry at room temperature (RT), after which they were fixed in 4% paraformaldehyde in PBS for 20 minutes, followed by washings in PBS and dehydration in 30%, 50%, 75%, 95%, and 100% ethanol. They were then air dried. All the solutions used in the in situ hybridization with RNA probes were treated with diethylpyrocarbonate (DEPC, Fluka, Buch, Switzerland) at a final concentration of 0.1%.

PS PS PS PS

tivities of 3 x 1 06 cpm/40 [lI. Both hydrolyzed and unhydrolyzed probes were used for in situ hybridization. Alkaline hydrolysis using NaHCO3 buffer was used for the treatment of full-length transcripts.22

were

Preparation of Probes Riboprobe Gemini system (Promega) was applied to construct riboprobes of laminin and type IV collagen cDNAs.19 For this purpose, vectors pSP 64 and pSP 65 containing a promoter for SP 6 RNA polymerase were used. A 906 bp Pst fragment from HL2-coding for the BlI domain of the laminin Bi chain gene20 was cloned into the M 13 polylinker site of pSP 64. A 91 6-bp BamH I-Hind III fragment from HT 21 encoding the NC1 domain and part of the 3' untranslated region of the pro al (IV) gene21 was cloned into both pSP 64 and pSP 65 vectors. For transcription, Promegas Riboprobe transcription kit (Promega, Madison, WI) was used, and transcripts were labeled with 35S-rCTP, yielding specific ac-

In Situ Hybridization In situ hybridization with anti-sense and sense RNA probes followed with certain modifications the protocol described by Hogan et a122 and Hoeffler et al.23 Prehybridization steps included incubation in 0.2 mol/l (molar) HCI (20 minutes at RT), followed by 5 minutes' wash in DEPC-H20 and proteinase K treatment (1 mg/ml, 15 to 30 minutes, 37°C). After that the sections were dipped in 0.2% glycine in phosphate-buffered saline (PBS) for 30 seconds, washed twice in PBS for 30 seconds, and proteolysis was stopped by immersing the sections in 4% paraformaldehyde in PBS for 20 minutes. The sections were washed in PBS and acetylated in 0.25% to 0.5% acetic anhydride in 0.1 mol/l triethanolamine for 10 minutes. After 5 minutes' wash in PBS and dehydration, the sections were allowed to air dry for 1 to 2 hours at RT before placing the hybridization mixture on them. The mixture contained the radioactive RNA probe, 10 mmol/l (millimolar) dithiothreitol (DTT), 10 mmol/l TRIS-HCI, 10 mmol/I NaPO4, 5 mmol/l ethylenediaminetetra-acetic acid (EDTA), 0.3 mol/l NaCI, yeast tRNA (1 mg/ml), deionized formamide 50% of the volume and dextran sulphate 10% of the volume, 0.02% (wt/vol) Ficoll, 0.02% (wt/vol) polyvinylpyrrolidone, and bovine serum albumin (0.2 mg/ml). Before adding into the hybridization mixture, the probe was denatured by boiling for 1 minute and placed on ice. Of the hybridization mixture, 40 ,ul (approximately 7 x 105 cpm/ 10 ,ul hybridization buffer) was placed on each section, after which sections were covered with DEPC-H20washed coverslips. The hybridization took place at 500C overnight. After hybridization, the coverslips were removed in a washing buffer containing all the other constituents of hybridization mixture except for dex-

Laminin mRNA in Malignant Fibrous Histiocytoma

1063


tran sulfate and tRNA. This took place at 600C for 2 x 30 minutes.24 Slides then were washed in fresh buffer at 50° for 1 to 4 hours, rinsed in 0.5 mol/l NaCI in 10 mmol/l TRIS-HCI, 1 mmol/l EDTA (TE) at 370C for 15 minutes, and they then were incubated in 0.5 mol/l NaCI in TE containing 40 ,ug/ml RNAse A (Sigma Chemical Co., St. Louis, MO) at 37°C for 30 minutes. Washings proceeded as follows: 0.5 mol/l NaCI in TE (370C, 15 minutes), 2 x SSC (45°C, 2 x 15 minutes), 1 x SSC (450C, 2 x 15 minutes). After washings, the sections were dehydrated in graded series of ethanol containing 300 mmol/l ammonium acetate and air dried at RT for 1 to 2 hours. Autoradiography was performed by dipping the slides into Kodak NTB-3 nuclear track emulsion diluted 1:1 with 1% glycerol in H20. After exposing for 7 to 15 days, the slides were developed in Kodak Dl 9 developer at RT for 5 minutes, rinsed in acetic acid, and fixed for 5 minutes at RT. The sections were counterstained with hematoxylin and eosin.

Antibodies Against Laminin and Type IV Collagen Antibodies against the 7S domain of type IV collagen and the P1 fragment of laminin were gifts from Drs. Leila Risteli, MD, and Juha Risteli, MD, Collagen Research Unit, Department of Medical Biochemistry, University of Oulu, Finland. The 7S domain of type IV collagen was purified from human kidney,25 and the fragment P1 of laminin from human placenta.26 Antisera were raised in rabbits

and specific antibodies were prepared by immunoadsorption on the relevant antigen, coupled to Sepharose 4B, after cross-adsorption with other immobilized extracellular matrix proteins. Sternberger's peroxidase-antiperoxidase procedure27 was used on sections cut from formalin-fixed paraffin-embedded specimens. The sections were treated with 0.4% pepsin (Merck, Rahway, NJ) in 0.01 mol/l HCI to enhance the availability of antigenic determinants.28 Before antigen-antibody reaction, the endogenous peroxidase was inactivated with 0.1% hydrogen peroxide in methanol. For control stainings, phosphate-buffered saline (PBS) and normal rabbit serum were used instead of the primary antibody.

Results Results of In Situ Hybridization The results are shown in Table 2. Evidence of laminin synthesis could be observed at the mRNA level in 11 MFHs. Ten of these were of the pleomorphic-storiform subtype and one was of the giant cell subtype. Signals for laminin mRNA could clearly be ascribed to the neoplastic cells and to the endothelial cells of the capillaries, whereas there were no signals seen in the small reactive histiocytelike cells among the neoplastic tumor cells (Figures 1 and 2A). Among the tumor cells, the pleomorphic giant cells were also usually positive (Figure 1). The osteoclastlike giant cells in the

Table 2. Presence of Laminin and Type IV Collagen in Tumor Cells and Endothelial Cells in MFH In situ hybridization Case Subtype Immunohistochemistry Laminin TIVC Laminin TIVC mRNA mRNA immunoreactivity Tumor cells/endothelial cells Intracellular/extracellular fibers

GC PS PS PS PS PS PS PS PS IS PS PS PS PS

1 2 3 4 5 6 7 8 9 10 11 12 13 14

+ ++

-+ +

+I+

+/- + + -+ +-

+

+--I+I-

-+ -I-I-

-I-

-I-

+1+ -I-1-I-I+ +/+

-1+ -I-1-I-I+I+

+

+ ++ --

*

+/+ + +I+

+/+ +- + + +I+ + --I+/- + +

-/+

-I+

-1+

-1+

-I+

-I+ -I+-

+/+ +I+

+-I+ +

Not done + + = strongly positive + = positive - = negative

*

+- =

TIVC

=

inconclusive type IV collagen; GC

=

giant-cell subtype; PS

=

pleomorphic-storiform subtype; IS

=

inflammatory subtype.

-I-

1064


AJP November 1991, Vol. 1-39, No. 5

Figure 1. 7be tumor cells in this case of a pleomorphic storiform subtype of MFH (case 8) shou' strong signalsfor laminin mRNA. Signals can also be observed in the giant pleomorphic tumor cells (arrow), In situ hybridization, X250.

giant cell variant remained negative, however (Figure 3A). No convincing signals for type IV collagen mRNA could be detected in the tumor cells of any MFHs, but in most cases the endothelial cells of the vascular channels in the tumor tissue showed strong signals for type IV collagen mRNA (Figure 3B). These were much stronger than the corresponding signals for the laminin mRNA. Because of the abundance of developing vascular structures in MFHs, it was sometimes difficult to evaluate whether the signals for type IV collagen mRNA were in immature endothelial cells or in neoplastic cells. In one case of MFH, the epidermis overlying the tumor could be seen in the histologic sections. Here strong signals for both laminin and type IV collagen mRNA could be observed in the basal cells of the epidermis.

Results of Immunohistochemical Stainings Nine tumors were immunohistochemically positive for laminin, eight of the pleomorphic-storiform subtype and one of the giant cell subtype. The laminin positivity in these cases was clearly intracytoplasmic and located in the atypical neoplastic cells. Occasional weak extracellular positivity could be observed only in three cases, all of which also expressed intracellular positivity for laminin. Except for one case of the pleomorphic storiform subtype, there was no intracellular positivity for type IV collagen in any of the MFHs, but there were three cases in which occasional weak extracellular positivity could be observed, the same ones that also expressed extracel-

lular positivity for laminin. One of these was also the only MFH with intracellular type IV collagen reactivity. The BMs of the vascular channels in the tumor tissue stained linearly for both laminin and type IV collagen.

In Situ Hybridization Results Compared with Immunohistochemistry Comparison of the in situ hybridization results with the immunohistochemical findings showed that all the MFHs with immunohistochemically detectable laminin also possessed signals for laminin mRNA, but there were two MFHs in which laminin could not be detected immunohistochemically even though the in situ hybridization method gave positive results. With respect to type IV collagen, the three cases that were immunohistochemically positive could not be shown to contain the corresponding mRNA in the tumor cells. The only cells that convincingly contained type IV collagen mRNA were the endothelial cells of vessels that also had linear BMs in the corresponding immunohistochemical stainings.

Discussion The results show that MFH tumor cells are capable of synthesizing mRNA specific to laminin. This is in accordance with our previous immunohistochemical report, in which intracytoplasmic laminin reactivity could be detected in these cells.15 Positive laminin immunoreactivity


1065


Figure 2. A: Fusiform neoplastic tumor cells of a pleomorphic-storzform subype of MFH (case 6) show strong signals for laminin mRNA (arrows), In situ hybridization, X250. B: Immunohistochemical positivity for laminin can be observed intracellularly in the tumor cells (arrows), Peroxidase-antiperoxidase, x250.

was verified also in our present investigation. The clear

correlation between the results obtained by immunohistochemistry and by in situ hybridization (Table 2) confirms that the intracytoplasmic laminin reactivity is indeed the result of synthetic activity and not derived from phagocytic activity in the tumor cells. There were two cases, however, in which positivity for laminin could not be seen immunohistochemically even though signals for laminin mRNA could be detected in in situ hybridization (see Table 2). The reason for this may reside in the poorer sensitivity of the former method. Con-

versely, the result may reflect defective translation of mRNA molecules, resulting in an altered molecular structure in laminin. This may lead to a situation in which antibodies directed toward the P1 fragment of laminin are not able to recognize their specific laminin epitope. We were unable to demonstrate any convincing signals for type IV collagen mRNA in the neoplastic cells of the MFHs, but there were strong signals in the endothelial cells of the developing or reactive capillaries within the tumor tissue. The signals for type IV collagen mRNA in these cells also were shown to be more abundant than

1066



Figure 3. A: The fusiform neoplastic tumor cells of the giant cell subtype of MFH (case 1) give strong signals for laminin mRNA while the osteoclastlike giant cells (arrows) are negative. In situ hybridization, X250. B: Tumor cells of the giant cell subtype of MFH contain no signalsfor type IVcollagen mRNA either in the neoplasticfusiform cells or the osteoclastlike giant cells. Conversely, the endothelial cells of the vascular channels in the tumor tissue give strong signals for type IV collagen mRNA (arrows), In situ hybridization, X250.

the corresponding signals for laminin mRNA. It has been shown previously by in situ hybridization that the endothelial cells of developing capillaries in the early placenta contain type IV collagen mRNA more abundantly than laminin mRNA.19 A similar finding thus could also be observed in the tumor capillaries of MFH. The complex immunohistochemical spectrum of MFH, with reactivities to several differentiation antigens, is thought to be due to the capability of the primitive mesenchymal cells, the putative ancestor cells of MFHs, to express several mesenchymal differentiation lineages.11

The synthetic activity of laminin observed in the majority of these tumors is in accordance with this hypothesis. Although laminin is synthesized by BM-producing cells such as smooth or striated muscle cells, Schwann cells, or lipocytes, the synthesis of this molecule by MFH cells could be regarded as a sign of differentiation toward such mesenchymal nonhistiocytic cell lineages. In fact, a couple of recent immunohistochemical investigations even report cytokeratin, desmin, and neurofilament positivity in some cases of MFH.67 Fibroblasts, which are normally not invested with a


1067


BM, have also been shown to be able to synthesize laminin,"9 2and thus the laminin synthesis observed in MFH tumor cells does not rule out the possibility of fibroblastic differentiation in these neoplasms. This is also implied by some other investigations describing fibroblastic differentiation in MFH, either ultrastructurally or by enzyme histochemistry.3'4 There is no exact information available on the capability of histiomonocytic cells to synthesize laminin, but it has been reported that at least macrophages possess laminin receptors on their cell membranes.30 We could not detect any laminin-specific mRNA here in the nonneoplastic histiocytic cells, which are known to be present in substantial numbers among the tumor cells of MFH.8'11 There are, however, previous immunohistochemical investigations that report reactivity of antibodies to monoclonal histiomonocytic cells in a minority of MFHs.9'11 Even though the existence of laminin mRNA in tumor cells may be considered to contradict histiomonocytic cell differentiation, the possibility of the existence of a subgroup of MFH with histiomonocytic differentiation cannot be totally excluded. It is interesting to note that in spite of pronounced laminin synthesis, the tumor cells of MFH do not possess an extracellular BM around them.5'3' Also extracellular fibers containing BM proteins are scarce in MFH.15 A corresponding situation has been described in cases of anaplastic carcinomas, in which the tumor cells contain intracytoplasmic immunoreactivity indicative of laminin and type IV collagen but lack extracellular BMs.32'33 Defective BM formation of this kind has been attributed to the activity of different types of BM-degrading collagenases and proteases that tumor cells produce.336 Conversely, the defective BM-forming capacity could be due to the insufficient extracellular assembly of BM-forming molecules.36 In MFH, this insufficiency may be due partly to the meager or nonexistent production of type IV collagen by tumor cells. Because of the widespread laminin-synthesizing capacity of mesenchymal cells and tumors, laminin cannot be used as a differentiation marker between any of these tumors. In a way the situation is analogous to the use of alpha-i-antichymotrypsin and alpha-i-antitrypsin as differentiation markers. These markers, which previously were considered to indicate fibrohistiocytic differentiation and were used to substantiate the diagnosis of MFH, have been immunohistochemically shown to be present in a wide variety of tumors.37'38 Consequently their value as fibrohistiocytic markers has been questioned.37'38 The in situ hybridization method could be valuable in resolving whether the immunoreactivity for alpha-iantichymotrypsin or alpha-i-antitrypsin in tumors is due to the sythetic activity or to uptake of these proteins by the tumor cells.

Acknowledgments The authors thank Ms. Heli Auno, Ms. Annikki Huhtela, and Ms. Mirja Vahera for expert technical assistance, and Mr. Tapio Leinonen for preparing the micrographs. We also thank Professor Veli-Pekka Lehto, Head of the Department, for his comments.

References 1. Enzinger F, Weiss S: Soft tissue tumors. St Louis, CV Mosby, 1988, pp 223-300 2. Ozzello L, Stout AP, Murray MR: Cultural characteristics of malignant histiocytomas and fibrous xanthomas. Cancer 1963,16:331-344 3. Hoffman MA, Dickersin GR: Malignant fibrous histiocytoma: An ultrastructural study of eleven cases. Hum Pathol 1983, 14:913-922 4. Wood GS, Beckstead JH, Turner RR, Hendrickson MR, Kempson RL, Warnke RA: Malignant fibrous histiocytoma tumor cells resemble fibroblasts. Am J Surg Pathol 1986,

10(5):323-335 5. Fu Y-S, Gabbiani G, Kaye G, Lattes R: Malignant soft tissue tumors of probable histiocytic origin (malignant fibrous histiocytomas): General considerations and electron microscopic and tissue culture studies. Cancer 1975,35:176-198 6. Lawson CW, Fisher C, Gatter KC: An immunohistochemical study of differentiation in malignant fibrous histiocytoma. Histopathology 1987, 11:375-387 7. Miettinen M, Soini Y: Malignant fibrous histiocytoma. Heterogeneous patterns of intermediate filament proteins by immunohistochemistry. Arch Pathol Lab Med 1989,113:13631366 8. Brecher ME, Franklin WA: Absence of mononuclear phagocyte antigens in malignant fibrous histiocytoma. Am J Clin Pathol 1986, 86:344-348 9. Strauchen JA, Dimitriu-Bona A: Malignant fibrous histiocytoma. Expression of monocyte/macrophage differentiation antigens detected with monoclonal antibodies. Am J Pathol 1986,124:303-309 10. Sieber SC, Lopez V, Rosai J, Buckley PJ: Primary tumor of spleen with morphologic features of malignant fibrous histiocytoma. Immunohistochemical evidence for a macrophage origin. Am J Surg Pathol 1990, 14(1 1):1061-1070 11. Soini Y, Miettinen M: Immunohistochemistry of markers of histiomonocytic cells in malignant fibrous histiocytomas: A monoclonal antibody study. Pathol Res Pract 1990,

186:759-767 12. Martinez-Hernandez A, Amenta PS: The basement membrane in pathology. Lab Invest 1983, 48:656-677 13. Miettinen M, Foidart J-M, Ekblom P: Immunohistochemical demonstration of laminin, the major glycoprotein of basement membranes, as an aid in the diagnosis of soft tissue tumors. Am J Clin Pathol 1983, 73:306-311 14. Ogawa K, Oguchi M, Yamabe H, Nakashima Y, Hamashima Y: Distribution of collagen type IV in soft tissue tumors. Cancer 1986, 58:269-277 15. Soini Y, Autio-Harmainen H, Miettinen M: Immunoreactivity

1068



for laminin and type IV collagen in malignant and benign fibrous histiocytoma. J Pathol 1989, 158:223-228 16. Sandberg M, Vuorio E: Localization of type 1, 11 and IlIl collagen mRNAs in developing human skeletal tissues by in situ hybridization. J Cell Biol 1987, 104:1077-1084 17. Sandberg M, Autio-Harmainen H, Vuorio E: Localization of the expression of types 1, III and IV collagen, TGF-B1 and c-fos genes in developing human calcarial bones. Dev Biol 1988,130:324-334 18. Syrjanen S, Syrjanen K: An improved in situ DNA hybridization protocol for detection of human papillomavirus (HPV) DNA sequences in paraffin-embedded biopsies. J Virol Methods 1986,14:293-304 19. Autio-Harmainen H, Sandberg M, Pihlajaniemi T, Vuorio E: Synthesis of laminin and type IV collagen by trophoblastic cells and fibroblastic stromal cells in the early human placenta. Lab Invest 1991, 64:483-491 20. Pikkarainen T, Eddy R, Fukushima Y, Byers M, Shows T, Pihlajaniemi T, Saraste M, Tryggvason K: Human laminin Bi chain. A multidomain protein with gene (LamBl) locus in the q22 region of chromosome 7. J Biol Chem 1987, 262:1 0454-1 0462 21. Pihlajaniemi T, Tryggvason K, Myers JC, Kurkinen M, Lebo R, Cheung MC, Procop DJ, Boyd CD: cDNA clones coding for the Proalphal (IV) chain of human type IV procollagen reveal an unusual homology of amino acid sequences in two halves of the carboxyterminal domain. J Biol Chem 1985, 260:7681-7687 22. Hogan B, Costantini F, Lacy E: Manipulating the mouse embryo. A laboratory Manual. Cold Spring Harbor, NY, Cold Spring Harbor Laboratory, 1968, pp 228-241 23. Hoeffler H, Childers H, Montminy MR, Lechan RM, Goodman RH, Wolfe HJ: In situ hybridization methods for detection of somatostatin mRNA in tissue sections using antisense RNA probes. Histochem J 1986, 18:597-604 24. Holland PWH, Harper SJ, McVey JH, Hogan BML: In vivo expression of mRNA for the Ca-binding protein SPARC (Osteonectin) revealed by in situ hybridization. J Cell Biol 1987, 105:473-482 25. Risteli J, Bachinger HP, Engel J, Furthmayr H, Timpl R: 7-S collagen: Characterization of an unusual basement membrane structure. Eur J Biochem 1980, 108:239-250

26. Risteli L, Timpl R: Isolation and characterization of pepsin fragments of laminin from human placenta and renal basement membranes. Biochem J 1981, 193:749-755 27. Sternberger LA: Immunocytochemistry. 2nd edition. New York, Wiley, 1979, pp 104-129 28. Ekblom P, Miettinen M, Rapola J, Foidart JM: Demonstration of laminin, a basement membrane glycoprotein in routine processed formalin fixed human tissues. Histochemistry 1982, 75:301-307 29. Woodley DT, Stanley JR, Reese MJ, O'Keefe EJ: Human dermal fibroblasts synthesize laminin. J Invest Dermatol 1988, 90:679-683 30. Huard TK, Malinoff HL, Wicha MS: Macrophages express a plasma membrane receptor for basement membrane laminin. Am J Pathol 1986,123:365-370 31. Taxy JB, Battifora H: Malignant fibrous histiocytoma. An electron microscopic study. Cancer 1977, 40:254-267 32. Albrechtsen R, Nielsen M, Wewer U, Engvall E, Ruoslahti E: Basement membrane changes in breast cancer detected by immunohistochemical staining for laminin. Cancer Res 1981, 41:5076-5081 33. Barsky H, Siegal GP, Janiotta F, Liotta LA: Loss of basement membrane components by invasive tumors but not by their benign counterparts. Lab Invest 1983, 49:140-147 34. Liotta LA, Abe S, Robey PG, Martin GR: Preferential digestion of basement membrane collagen by an enzyme derived from a metastatic murine tumor. Proc Natl Acad Sci 1979, 76:2268-2272 35. Shields SE, Ogilvie DJ, McKinnell RG, Tarin D: Degradation of basement membrane collagens by metalloproteases released by human, murine and amphibian tumors. J Pathol 1984,143:193-197 36. Bosman FT, Havenith M, Cleutjens JPM: Basement membranes in cancer. Ultrastruct Pathol 1985, 8:291-304 37. Soini Y, Miettinen M: Widespread immunoreactivity for alpha-i-antichymotrypsin in different types of tumors. Am J Clin Pathol 1988, 89:131-136 38. Soini Y, Miettinen M: Alpha-1 -antitrypsin and lysozyme. Their limited significance in fibrohistiocytic tumors. Am J Clin Pathol 1989, 91:515-521