0022-534 7/91/1462-0428$03.00/0 THE JOURNAL OF UROLOGY Copyright© 1991 by AMERICAN UROLOGICAL ASSOCIATION, INC.
Vol. 146, 428-432 August 1991 Printed in U.S.A.
TYPE IV PROCOLLAGEN mRNA REGULATION: EVIDENCE FOR EXTRACELLULAR MATRIX/CYTOSKELETON/NUCLEAR MATRIX INTERACTIONS IN HUMAN UROTHELIUM RICHARD N. SCHLUSSEL, MICHAEL J. DROLLER AND BRIAN C.-S. LIU* From the Department of Urology, Mount Sinai School of Medicine, New York, New York
ABSTRACT
The absence of basement membrane components correlates with tumor stage and progression in human bladder cancers. We have previously shown that invasive tumors possess the ability to degrade basement membrane. However, the presence of basement membrane may be affected not only by its degradation, but by its synthesis and deposition as well. Our results in the present study suggest that while the invasive human transitional carcinoma cell line EJ has an increased amount of type IV procollagen mRNA when compared to the non invasive RT4 cell line, type IV collagen staining is absent in the invasive EJ cells and intensely present in the non-invasive RT4 cells. Moreover, when EJ cells were grown on an artificial basement membrane (Matrigel), type IV procollagen mRNA expression was down-regulated to the levels seen with the non-invasive RT4 cells. We also discovered that the invasive cells, when grown on Matrigel, appeared morphologically different from the same cells grown on plastic tissue cultures. We conclude that a deficient basement membrane in invasive cancer cells may be due not only to active proteolytic activity but also to an abnormal production and deposition of extracellular matrix components. In addition, we also demonstrated that basement membrane components may have a significant effect on epithelial cell morphology and gene regulation, and that any alterations of the extracellular matrix-cytoskeleton-nuclear matrix interactions can lead to altered gene regulations and cell function. KEY WORDS: bladder, extracellular matrix, mRNA
Cancer cells, unlike normal cells, have the ability to produce a wide variety of cell populations with differing biological properties. This cellular variation or tumor cell heterogeneity, is demonstrated in a wide range of cellular characteristics including cell surface antigens, karyotypes, oncogenes and growth factors expression, growth properties, and the ability to invade and produce metastases. 1 Human bladder tumors reflect such heterogeneity clinically: some behave in a benign fashion, and others are highly aggressive and lead rapidly to metastatic disease and death. 2 What is not yet resolved is the mechanisms responsible for this state of cellular heterogeneity. It is hypothesized that cellular heterogeneity is due, in part, to genetic changes of a tumor cell through DNA instability within the genome during tumor progression. This process of DNA instability has been suggested to provide the tumor with varying phenotypic clones of cells. 3 Nowell proposed that the original tumor may develop from a single cell, but as proliferation of the cells continued, progeny cells became unstable and a change occurred in the DNA that was transmitted to subsequent daughter cells. If this genetic instability continued, it soon would produce a wide variety of different cell types within a single tumor. 3 The amount of DNA or chromosomal content of a tumor cell can be altered with tumor progression, often resulting in a state of DNA polyploidy or the more variable aneuploidy. This genetic instability, reflected in the DNA content, may lead to the alteration of tumor cell properties such as drug sensitivity, metabolism, metastatic potential, and growth rates.3 Tumor metastasis begins with the growth of tumor cells and Accepted for publication February 1, 1991. *Requests for reprints: Department of Urology, Mt. Sinai School of Medicine, One Gustave Levy Place, New York NY 10029. Supported by Grants from the New York Academy of Medicine, the Edwin Beer Award, and the National Cancer Institute (lROl CA51968).
invasion into host stroma surrounding the primary neoplasm. The absence of intact basement membrane components corre lates with tumor stage and progression.4 Invasive bladder can cers have been shown to have deficient basement membranes when compared to their non-invasive and normal counterparts.4 This was attributed to increased enzymatic degradation of the basement membrane.5 Indeed, Weiss et al.6' 7 demonstrated that the absence of basement membrane laminin in invasive human transitional cell carcinoma (TCC) cells was due to the enhanced production and the redistribution of activated protease cathep sin B to the plasma membrane of the invasive cells. However, the presence of basement membrane may be af fected not only by its degradation, but by its synthesis and deposition as well. Because we believed that the extracellular matrix is not only a static barrier but a dynamic regulator of the adjacent epithelium, we studied the effects of invasive bladder cancer cells grown on an artificial basement membrane in regards to gene expression and cellular morphology. In this paper, we report that the expression of basement membrane component mRNA is regulated by the deposition of the extra cellular matrix. Our findings suggest that the basement mem brane has a stabilizing effect on nuclear function and that the absence of basement membrane in invasive cells may affect their gene expression, resulting in genetic instability. MATERIALS AND METHODS
Cell lines and culture conditions. Human transitional cell carcinoma cell lines EJ and RT4 were used in this study. RT48 is a cell line that originated from a 63-yr.-old patient who underwent a cystectomy because of recurrent multifocal super ficial, low-grade papillary TCC. EJ 9 is a grade III, invasive TCC cell line that was derived from a patient undergoing cystectomy for invasive bladder cancer. All cultures were maintained in RPMI-1640 containing 10% fetal calf serum, two mM gluta-
428
429
TYPE IV PROCOLLAGEN MRNA REGULATION
8 FIG. 1. A, immunocytochemical analysis of non-invasive RT4 hu man bladder cancer cells grown on plastic culture flasks, stained with type IV collagen monoclonal antibodies and counterstained with he matoxylin. Note intense intracellular and extracellular staining (lOOX). B, immunocytochemical analysis of invasive EJ human bladder cancer cells grown on plastic culture flasks, stained with type IV collagen monoclonal antibodies and counterstained with hematoxylin. Note absence of intracellular or extracellular staining (lOOX).
mine, 1% non-essential amino acids, 100 µg. penicillin/ml., and 100 µg. streptomycin/ml. 108 cells grown on tissue culture flasks were removed and used for RNA isolation (see below). Extracellular matrix or Matrigel (Collaborative Research) containing basement membrane components such as type IV collagen, laminin, entactin, and heparin sulfate glycoproteins, was diluted 1:2 and coated onto culture dishes at 5 ml./75 cm.2 dish and preincubated for five hr. in a humidified incubator at 37C before using as described by the manufacturer. 10 7 log phase EJ cells were plated onto the matrix, and an additional 10 ml. of culture media was added to the flask. After 72 hr., the cells were removed from the matrix with limited collagenase digestion as described, 10 centrifuged, and total RNA was im mediately isolated for analysis. Northern analysis of mRNA. RNA was prepared by the guanidinium isothiocyanate method,11 separated in denaturing formaldehyde gels,12 and transferred to nylon membranes for hybridization.13 Blots containing 20 µg. total RNA were probed for type IV procollagen sequences. Probe was labeled with 32P dCTP (Amersham Corp.) with the use of the random-priming method.14 Type IV procollagen cDNA probe was generously provided by Dr. C. D. Boyd of Robert Wood Johnson Medical School.15 After two hr. of prehybridization, hybridization was carried out overnight at 42C in the presence of 2X SSC (SSC
FIG. 2. A, photomicrograph of invasive EJ cells grown on plastic culture flasks. Note that cells have spindled, tentacular appearance (lOOx). B, photomicrograph of invasive EJ cells grown on artificial basement membrane Matrigel. Note formation of small cell clusters resembling papillary-like projections as seen in vivo (lOOX).
is 150 mM NaCl, 1.5 mM sodium citrate), one µg./ml. BSA, one µg./ml. Ficoll, one µg./ml. polyvinylpyrollidine, 0.1% SDS, 25 mM NaH2PO., 50% deionized formamide, 100 µg.jml. sheared
salmon sperm DNA, and 3-8 x 106 cpm of type IV procollagen cDNA probe. Blots were washed first in 2X SSC at room temperature, followed by 0.1% SDS and 0.2X SSC at 45C, and finally with O.lx SDS and 0.2X SSC at 60C. Blots were then exposed to XAR-5 film (Eastman Kodak Co.) with intensifying screen overnight at -70C as described.12 Immunocytochemistry. Cells were removed from non-matrix coated tissue culture flasks with the use of sterile cell scrapers (Falcon). Cells were centrifuged onto slides with the use of a Cytospin (American Scientific Products, Inc.). The cells were then fixed in sequential dilutions of alcohol, and subsequently in phosphate buffered saline (PBS). To detect the presence or absence of type IV collagen, the cells were incubated in 10% pepsin at 37C for one hour to unmask the epitope. The cells were incubated in 3% H20 2 to block endogenous peroxidase activity, blocked with normal serum for 20 minutes, and then incubated with mouse monoclonal anti-human type IV collagen antibodies (Miles Laboratory) (1:150) overnight. The cells were subsequently incubated with biotinylated rabbit anti-mouse antibody (Vector Lab) (1:200) for 30 minutes, and followed with the avidin-biotin peroxidase complex for 20 minutes, and then reacted with 3,3' -diaminobenzidene. The cells were then lightly counterstained with hematoxylin.
1
SCHLUSSEL, DROLLER AND LIU
430
EJ / ECM
RT EJ
7kb
7kb 5kb
5kb
2kb
2 kb
FIG. 3. Northern analysis for type IV procollagen mRNA in invasive EJ and non-invasive RT4 human bladder cancer cell lines grown on plastic culture flasks. Note band at 7 kb in EJ lane, representing increased type IV procollagen mRNA. Bands seen at 5 kb and 2 kb represent non-specific binding to ribosomal RNA (288 and 188 respec tively) and act as internal control.
FIG. 4. Northern analysis for type IV procollagen mRNA in invasive EJ human bladder cancer cells grown on artificial basement membrane Matrigel. Note absence of 7 kb band (even when autoradiograph was deliberately over-exposed) which was present in EJ cells grown on plastic surfaces (figure 3). Bands seen at 5 kb and 2 kb represent non specific binding to ribosomal RNA (288 and 188 respectively) and act as internal control.
RESULTS
basement membrane Matrigel, the expression of the type IV procollagen mRNA was down-regulated to the levels seen with the non-invasive RT4 cells (figure 4). Even when the autora diographs were deliberately over-exposed, no type IV procolla gen mRNA was detected (figure 4). In all cases, the 28S and the 18S ribosomal RNA bands were of the same intensity and served as an internal control for RNA degradation, amounts, and hybridization efficiency (figures 3 and 4).
Immunocytochemical detection of type IV collagen. The non
invasive RT4 cells, when grown on plastic, exhibited intense staining for type IV collagen both intra- and extracellularly on immunocytochemical studies (figure lA). The invasive EJ cells grown on plastic displayed none of the staining for type IV collagen, neither intracellularly or extracellularly (figure lB). Morphology of cells grown on extracellular matrix or on plastic.
The invasive EJ cells when grown on plastic displayed a spin dle-shaped, tentacular appearance (figure 2A). However the same cells when grown on an artificial basement membrane Matrigel, formed small clusters resembling papillary-like pro jections as seen in vivo (figure 2B). Expression of type IV procollagen mRNA. There was a marked expression of type IV procollagen mRNA at seven kb for the invasive EJ cells when the cells were grown on plastic. However, the noninvasive RT4 cells failed to reveal any significant type IV procollagen mRNA (figure 3). When the invasive EJ cells were grown on the artificial
DISCUSSION
For a tumor to be invasive, the cells must penetrate the basement membrane and then move into the surrounding stroma.5 Previously, the process of extracellular matrix degra dation was believed to be the main explanation for the absence of intact basement membranes in invasive tumors.5 However, the presence of basement membrane is in a state of constant flux. We have demonstrated that the undifferentiated invasive cancer cells has difficulty in synthesizing basement membrane
TYPE I'! PROCOLLAGEN MRNA REGl-)LATION
140kd Flbrnneclln binding prolelna
Collagen
FIG. 5. Schematic of extracellular matrix/cytoskeleton/nuclear ma trix interactions. Expanded portion is schematic of cytoplasmic mem brane interaction with extracellular matrix via integrin receptors (see text).
components as demonstrated by the absence of both intra- and extracellular type IV collagen in the invasive EJ cells (figure lB). The more differentiated non-invasive RT4 cells demon strated the presence of both intra- and extracellular type IV collagen (figure lA). These results are consistent with those obtained previously from our laboratory using an in vivo animal model in which the RT4 or EJ cells were inoculated per urethra into the bladder of nude mice.6 Immunohistochemical analysis of the implanted tumors revealed that the RT4 cells were surrounded by intact extracellular matrix, while the invasive EJ cells lacked extracellular matrix. 6 In our present study, we also demonstrated the effect on epithelial cell gene expression by an intact basement mem brane. When the invasive EJ cells were grown on plastic, because the cells were not capable of synthesizing the type IV collagen, the type IV procollagen message appeared to be up regulated (figure 3). However, when the invasive EJ cells were grown on an artificial basement membrane Matrigel, the type IV procollagen message was down-regulated (figure 4). The lack of significant hybridization to the type IV procollagen mRNA in the non-invasive RT4 cells can be explained by their ability to synthesize and deposit type IV collagen and laminin (figure lA). This observation is consistent with our hypothesis that the basement membrane may play a functional role in gene regulation. Others have shown that extracellular matrices have an affect, positively and dramatically, on the functional secretion of milk ·proteins in cultured mammary epithelial cells. 16 Streuli and BisseP7 have further shown that the levels of laminin, type IV collagen, and fibronectin mRNAs are highest in cells cultured on plastic surfaces and that the interaction between mammary
431
epithelial cells and both basement membrane and stromally derived matrices appears to exert a negative influence on the expression of mRNA for extracellular matrix components. We also found that the morphology of the EJ cell line also differed in the two environments. On plastic, the cells were spindle-shaped with a suggestion of pseudopodia-like projec tions (figure 2A). On Matrigel, the cells were clustered and had a papillary-like organization (figure 2B). One possible mechanism explaining these observations is based on known interconnections between the extracellular matrix, cytoskeleton, and nuclear matrix. 18 Coffey has proposed that the nuclear matrix actually forms an interlocking cellular matrix network by direct connection with the cytoskeletal matrix, and extends to the extracellular matrix. 18 The nuclear matrix is comprised of nuclear proteins called lamins, which only recently have been identified as members of the intermediate filament family.18 In addition to interac tions with chromosomes, the lamins also form a scaffold to which the nuclear pores are attached. Outside the nucleus, the cytoplasmic intermediate filaments are frequently close to the nuclear envelope. The cytoplasmic intermediate filaments may interact with the lamins, either directly via the nuclear pores or indirectly via a component of the nuclear envelope 18 (figure 5). The nuclear matrix is associated with RNA synthesis. Tran scriptional complexes have been identified on the nuclear ma trix. 19 O'Malley et al.20 observed that 95% of the unprocessed mRNA precursor for ovalbumin was associated with the nuclear matrix. When the intrans were processed out, the mature mRNA was released from the nuclear matrix. This suggests that the nuclear matrix was involved in mRNA processing and that any alterations in the nuclear matrix structures, which may occur as a result from the absence of an extracellular matrix, could alter important steps in transcription and RNA processing. Furthermore, Reid and her colleagues21 ' 22 demonstrated that the regulation of insulin messenger RNA abundance in cultured hepatocytes was dependent on the type of the extracellular matrix components on which the cells were grown. Therefore, the extracellular matrix can alter gene expression of cells in a specific manner. Not only are individual extracellular matrix components important but that the interaction of several of the components are involved in cell signaling and function. More recently, Fearon et al.23 identified a colorectal cancer gene which was deleted or altered in more than 70% of colorec tal cancers. This gene was found to specify a protein with sequences similar to those of the neural cell adhesion molecules. A gene altered in colorectal tumors which is similar in sequence to genes involved in cell-surface interactions is relevant to our present study of cell-basement membrane interactions. The extracellular matrix components contact transmembrane re ceptors through the cell membrane which includes the families of integrins and cell adhesion molecules. 24 These receptors contain sequences for phosphorylation in their cytoplasmic domain and may be important in providing signals that deter mine the location, polarity, and shape of cells. Signals from the extracellular matrix, therefore, may exert as much control over the behavior of cells as do hormones and other soluble mediators. Our results demonstrated that the basement membrane has a regulatory effect on both cellular morphology and nuclear function on urothelial cells. We also suggest that the absence of basement membrane associated with invasive human bladder tumor cells may not only facilitate their ability to penetrate into blood and lymphatic vessels, but may affect their gene expressions, contributing to further ge netic instability and leading to phenotypic variations which are the hallmark of malignancy. We are currently pursuing the molecular relationships between the extracellular matrix and the nuclear matrix. Results from our studies may elucidate the mechanisms responsible for tumor heterogeneity, as well as
432
SCHLUSSEL, DROLLER AND LIU
developmental anomalies such as Alport syndrome in which components of the extracellular matrix are also altered. Acknowledgments. The authors thank the assistance and friendship provided by Dr. Gregory Elder of the Center For Molecular Biology at Mt. Sinai School of Medicine. REFERENCES 1. Fidler, I. J. and Poste, G.: The cellular heterogeneity of malignant neoplasms: implications for adjuvant chemotherapy. Semin. On col., 12: 207, 1985. 2. Droller, M. J.: Transitional cell cancer: upper tracts and bladder. In: Campbell's Urology, ed. 5. Edited by P. C. Walsh, R. F. Gittes, A. D. Perlmutter, and T. A. Stamey. Vol. 2, pp. 13531366, Philadelphia: W.B. Saunders Co., 1986. 3. Nowell, P.: Mechanisms of tumor progression. Cancer Res., 46: 2203, 1986. 4. Conn, I. G., Crocker, J. and Wallace, D. M. A. et al.: Basement membranes in urothelial carcinoma. Br. J. Urol., 60: 536, 1987. 5. Liotta, L. A.: Tumor invasion and metastases: role of the basement membrane. Am. J. Pathol., 117: 339, 1984. 6. Weiss, R. E., Liu, B. C. S. and Ahlering, T. E. et al.: Mechanisms of human bladder tumor invasion: role of protease cathepsin B. J. Urol., 144: 798, 1990. 7. Redwood, S. M., Weiss, R. E., Droller, M. J. and Liu, B. C. S.: Abrogation of invasion by human bladder tumor cells using protease inhibitors. Surgical Forum, 41: 694, 1990. 8. Rigby, C. C. and Franks, L. M.: A human tissue culture cell line from a transitional cell tumor of the urinary bladder: growth, chromosome pattern, and ultrastructure. Br. J. Cancer, 24: 746, 1970. 9. Shih, C. and Weinberg, R. A.: Isolation of a transforming sequence from a human bladder carcinoma cell line. Cell, 29: 161, 1982. 10. Li, M. L., Aggeler, J. and Farson, D. A. et al.: Influence of a reconstituted basement membrane and its components on casein gene expression and secretion in mouse mammary epithelial cells. Proc. Natl. Acad. Sci. USA, 84: 136, 1987.
11. Chirgwin, J. M., Przybyla, A. L. and MacDonald, R. J. et al.: Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18: 5294, 1979. 12. Maniatis, T. E., Fritsch, E. F. and Sambrook, J.: In: Molecular Cloning: A Laboratory Manual, p. 545. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. 13. Fourney, R. M., Miyakoshi, J. and Day-III, R. S. et al.: Northern blotting: efficient RNA staining and transfer. Focus, 10: 5, 1987. 14. Feinberg, A. P. and Vogelstein, B. V.: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activ ity. Anal. Biochem., 137: 266, 1984. 15. Pihlajaniemi, T., Tyggvason, K. and Myers, J. C. et al.: cDNA clones for the pro-alpha 1 (IV) chain of human type IV procol lagen reveal an unusual homology of amino acid sequences in two halves of the carboxyl-terminal domain. J. Biol. Chem., 260: 7681, 1985. 16. Barcellos-Hoff, M. H., Aggeler, J. and Ram, T. G. et al.: Functional differentiation and alveolar morphogenesis of primary mammary epithelial cells cultured on reconstituted basement membrane. Development, 105: 223, 1989. 17. Streuli, C. H. and Bissell, M. J.: Expression of extracellular matrix components is regulated by substratum. J. Cell Biol., 110: 1405, 1990. 18. Nelson, W. G., Pienta, K. J., Barrack, E. R. and Coffey, D. S.: The role of the nuclear matrix in the organization and function of DNA. Ann. Rev. Biophys. Biophys. Chem., 15: 457, 1986. 19. Jackson, D. A., McCready, S. J. and Cook, P. R.: RNA is synthe sized at the nuclear cage. Nature, 292: 552, 1981. 20. Ciejek, E. M., Nordstromn, J. L., Tsai, M-J. and O'Malley, B. W.: Ribonucleic acid precursors are associated with the chick oviduct nuclear matrix. Biochemistry, 21: 4945, 1982. 21. Muschel, R., Khoury, G. and Reid, L. M.: Regulation of insulin mRNA abundance and adenylation: dependence on hormones and matrix substrate. Mo!. Cell Bio., 6: 337, 1986. 22. Reid, L. M.: Stem cell biology, hormone/matrix synergies and liver differentiation. Curr. Opin. Cell Biol., 2: 121, 1990. 23. Fearon, E. R., Cho, K. R. and Nigro, J. M. et al.: Identification of a chromosome 18q gene that is altered in colorectal cancers. Science, 247: 49, 1990. 24. Ruoslahti, E. and Pierschbacher, M. D.: New perspectives in cell adhesion: RGD and integrins. Science, 238: 491, 1987.
Vol. 146, 428-432 August 1991 Printed in U.S.A.
TYPE IV PROCOLLAGEN mRNA REGULATION: EVIDENCE FOR EXTRACELLULAR MATRIX/CYTOSKELETON/NUCLEAR MATRIX INTERACTIONS IN HUMAN UROTHELIUM RICHARD N. SCHLUSSEL, MICHAEL J. DROLLER AND BRIAN C.-S. LIU* From the Department of Urology, Mount Sinai School of Medicine, New York, New York
ABSTRACT
The absence of basement membrane components correlates with tumor stage and progression in human bladder cancers. We have previously shown that invasive tumors possess the ability to degrade basement membrane. However, the presence of basement membrane may be affected not only by its degradation, but by its synthesis and deposition as well. Our results in the present study suggest that while the invasive human transitional carcinoma cell line EJ has an increased amount of type IV procollagen mRNA when compared to the non invasive RT4 cell line, type IV collagen staining is absent in the invasive EJ cells and intensely present in the non-invasive RT4 cells. Moreover, when EJ cells were grown on an artificial basement membrane (Matrigel), type IV procollagen mRNA expression was down-regulated to the levels seen with the non-invasive RT4 cells. We also discovered that the invasive cells, when grown on Matrigel, appeared morphologically different from the same cells grown on plastic tissue cultures. We conclude that a deficient basement membrane in invasive cancer cells may be due not only to active proteolytic activity but also to an abnormal production and deposition of extracellular matrix components. In addition, we also demonstrated that basement membrane components may have a significant effect on epithelial cell morphology and gene regulation, and that any alterations of the extracellular matrix-cytoskeleton-nuclear matrix interactions can lead to altered gene regulations and cell function. KEY WORDS: bladder, extracellular matrix, mRNA
Cancer cells, unlike normal cells, have the ability to produce a wide variety of cell populations with differing biological properties. This cellular variation or tumor cell heterogeneity, is demonstrated in a wide range of cellular characteristics including cell surface antigens, karyotypes, oncogenes and growth factors expression, growth properties, and the ability to invade and produce metastases. 1 Human bladder tumors reflect such heterogeneity clinically: some behave in a benign fashion, and others are highly aggressive and lead rapidly to metastatic disease and death. 2 What is not yet resolved is the mechanisms responsible for this state of cellular heterogeneity. It is hypothesized that cellular heterogeneity is due, in part, to genetic changes of a tumor cell through DNA instability within the genome during tumor progression. This process of DNA instability has been suggested to provide the tumor with varying phenotypic clones of cells. 3 Nowell proposed that the original tumor may develop from a single cell, but as proliferation of the cells continued, progeny cells became unstable and a change occurred in the DNA that was transmitted to subsequent daughter cells. If this genetic instability continued, it soon would produce a wide variety of different cell types within a single tumor. 3 The amount of DNA or chromosomal content of a tumor cell can be altered with tumor progression, often resulting in a state of DNA polyploidy or the more variable aneuploidy. This genetic instability, reflected in the DNA content, may lead to the alteration of tumor cell properties such as drug sensitivity, metabolism, metastatic potential, and growth rates.3 Tumor metastasis begins with the growth of tumor cells and Accepted for publication February 1, 1991. *Requests for reprints: Department of Urology, Mt. Sinai School of Medicine, One Gustave Levy Place, New York NY 10029. Supported by Grants from the New York Academy of Medicine, the Edwin Beer Award, and the National Cancer Institute (lROl CA51968).
invasion into host stroma surrounding the primary neoplasm. The absence of intact basement membrane components corre lates with tumor stage and progression.4 Invasive bladder can cers have been shown to have deficient basement membranes when compared to their non-invasive and normal counterparts.4 This was attributed to increased enzymatic degradation of the basement membrane.5 Indeed, Weiss et al.6' 7 demonstrated that the absence of basement membrane laminin in invasive human transitional cell carcinoma (TCC) cells was due to the enhanced production and the redistribution of activated protease cathep sin B to the plasma membrane of the invasive cells. However, the presence of basement membrane may be af fected not only by its degradation, but by its synthesis and deposition as well. Because we believed that the extracellular matrix is not only a static barrier but a dynamic regulator of the adjacent epithelium, we studied the effects of invasive bladder cancer cells grown on an artificial basement membrane in regards to gene expression and cellular morphology. In this paper, we report that the expression of basement membrane component mRNA is regulated by the deposition of the extra cellular matrix. Our findings suggest that the basement mem brane has a stabilizing effect on nuclear function and that the absence of basement membrane in invasive cells may affect their gene expression, resulting in genetic instability. MATERIALS AND METHODS
Cell lines and culture conditions. Human transitional cell carcinoma cell lines EJ and RT4 were used in this study. RT48 is a cell line that originated from a 63-yr.-old patient who underwent a cystectomy because of recurrent multifocal super ficial, low-grade papillary TCC. EJ 9 is a grade III, invasive TCC cell line that was derived from a patient undergoing cystectomy for invasive bladder cancer. All cultures were maintained in RPMI-1640 containing 10% fetal calf serum, two mM gluta-
428
429
TYPE IV PROCOLLAGEN MRNA REGULATION
8 FIG. 1. A, immunocytochemical analysis of non-invasive RT4 hu man bladder cancer cells grown on plastic culture flasks, stained with type IV collagen monoclonal antibodies and counterstained with he matoxylin. Note intense intracellular and extracellular staining (lOOX). B, immunocytochemical analysis of invasive EJ human bladder cancer cells grown on plastic culture flasks, stained with type IV collagen monoclonal antibodies and counterstained with hematoxylin. Note absence of intracellular or extracellular staining (lOOX).
mine, 1% non-essential amino acids, 100 µg. penicillin/ml., and 100 µg. streptomycin/ml. 108 cells grown on tissue culture flasks were removed and used for RNA isolation (see below). Extracellular matrix or Matrigel (Collaborative Research) containing basement membrane components such as type IV collagen, laminin, entactin, and heparin sulfate glycoproteins, was diluted 1:2 and coated onto culture dishes at 5 ml./75 cm.2 dish and preincubated for five hr. in a humidified incubator at 37C before using as described by the manufacturer. 10 7 log phase EJ cells were plated onto the matrix, and an additional 10 ml. of culture media was added to the flask. After 72 hr., the cells were removed from the matrix with limited collagenase digestion as described, 10 centrifuged, and total RNA was im mediately isolated for analysis. Northern analysis of mRNA. RNA was prepared by the guanidinium isothiocyanate method,11 separated in denaturing formaldehyde gels,12 and transferred to nylon membranes for hybridization.13 Blots containing 20 µg. total RNA were probed for type IV procollagen sequences. Probe was labeled with 32P dCTP (Amersham Corp.) with the use of the random-priming method.14 Type IV procollagen cDNA probe was generously provided by Dr. C. D. Boyd of Robert Wood Johnson Medical School.15 After two hr. of prehybridization, hybridization was carried out overnight at 42C in the presence of 2X SSC (SSC
FIG. 2. A, photomicrograph of invasive EJ cells grown on plastic culture flasks. Note that cells have spindled, tentacular appearance (lOOx). B, photomicrograph of invasive EJ cells grown on artificial basement membrane Matrigel. Note formation of small cell clusters resembling papillary-like projections as seen in vivo (lOOX).
is 150 mM NaCl, 1.5 mM sodium citrate), one µg./ml. BSA, one µg./ml. Ficoll, one µg./ml. polyvinylpyrollidine, 0.1% SDS, 25 mM NaH2PO., 50% deionized formamide, 100 µg.jml. sheared
salmon sperm DNA, and 3-8 x 106 cpm of type IV procollagen cDNA probe. Blots were washed first in 2X SSC at room temperature, followed by 0.1% SDS and 0.2X SSC at 45C, and finally with O.lx SDS and 0.2X SSC at 60C. Blots were then exposed to XAR-5 film (Eastman Kodak Co.) with intensifying screen overnight at -70C as described.12 Immunocytochemistry. Cells were removed from non-matrix coated tissue culture flasks with the use of sterile cell scrapers (Falcon). Cells were centrifuged onto slides with the use of a Cytospin (American Scientific Products, Inc.). The cells were then fixed in sequential dilutions of alcohol, and subsequently in phosphate buffered saline (PBS). To detect the presence or absence of type IV collagen, the cells were incubated in 10% pepsin at 37C for one hour to unmask the epitope. The cells were incubated in 3% H20 2 to block endogenous peroxidase activity, blocked with normal serum for 20 minutes, and then incubated with mouse monoclonal anti-human type IV collagen antibodies (Miles Laboratory) (1:150) overnight. The cells were subsequently incubated with biotinylated rabbit anti-mouse antibody (Vector Lab) (1:200) for 30 minutes, and followed with the avidin-biotin peroxidase complex for 20 minutes, and then reacted with 3,3' -diaminobenzidene. The cells were then lightly counterstained with hematoxylin.
1
SCHLUSSEL, DROLLER AND LIU
430
EJ / ECM
RT EJ
7kb
7kb 5kb
5kb
2kb
2 kb
FIG. 3. Northern analysis for type IV procollagen mRNA in invasive EJ and non-invasive RT4 human bladder cancer cell lines grown on plastic culture flasks. Note band at 7 kb in EJ lane, representing increased type IV procollagen mRNA. Bands seen at 5 kb and 2 kb represent non-specific binding to ribosomal RNA (288 and 188 respec tively) and act as internal control.
FIG. 4. Northern analysis for type IV procollagen mRNA in invasive EJ human bladder cancer cells grown on artificial basement membrane Matrigel. Note absence of 7 kb band (even when autoradiograph was deliberately over-exposed) which was present in EJ cells grown on plastic surfaces (figure 3). Bands seen at 5 kb and 2 kb represent non specific binding to ribosomal RNA (288 and 188 respectively) and act as internal control.
RESULTS
basement membrane Matrigel, the expression of the type IV procollagen mRNA was down-regulated to the levels seen with the non-invasive RT4 cells (figure 4). Even when the autora diographs were deliberately over-exposed, no type IV procolla gen mRNA was detected (figure 4). In all cases, the 28S and the 18S ribosomal RNA bands were of the same intensity and served as an internal control for RNA degradation, amounts, and hybridization efficiency (figures 3 and 4).
Immunocytochemical detection of type IV collagen. The non
invasive RT4 cells, when grown on plastic, exhibited intense staining for type IV collagen both intra- and extracellularly on immunocytochemical studies (figure lA). The invasive EJ cells grown on plastic displayed none of the staining for type IV collagen, neither intracellularly or extracellularly (figure lB). Morphology of cells grown on extracellular matrix or on plastic.
The invasive EJ cells when grown on plastic displayed a spin dle-shaped, tentacular appearance (figure 2A). However the same cells when grown on an artificial basement membrane Matrigel, formed small clusters resembling papillary-like pro jections as seen in vivo (figure 2B). Expression of type IV procollagen mRNA. There was a marked expression of type IV procollagen mRNA at seven kb for the invasive EJ cells when the cells were grown on plastic. However, the noninvasive RT4 cells failed to reveal any significant type IV procollagen mRNA (figure 3). When the invasive EJ cells were grown on the artificial
DISCUSSION
For a tumor to be invasive, the cells must penetrate the basement membrane and then move into the surrounding stroma.5 Previously, the process of extracellular matrix degra dation was believed to be the main explanation for the absence of intact basement membranes in invasive tumors.5 However, the presence of basement membrane is in a state of constant flux. We have demonstrated that the undifferentiated invasive cancer cells has difficulty in synthesizing basement membrane
TYPE I'! PROCOLLAGEN MRNA REGl-)LATION
140kd Flbrnneclln binding prolelna
Collagen
FIG. 5. Schematic of extracellular matrix/cytoskeleton/nuclear ma trix interactions. Expanded portion is schematic of cytoplasmic mem brane interaction with extracellular matrix via integrin receptors (see text).
components as demonstrated by the absence of both intra- and extracellular type IV collagen in the invasive EJ cells (figure lB). The more differentiated non-invasive RT4 cells demon strated the presence of both intra- and extracellular type IV collagen (figure lA). These results are consistent with those obtained previously from our laboratory using an in vivo animal model in which the RT4 or EJ cells were inoculated per urethra into the bladder of nude mice.6 Immunohistochemical analysis of the implanted tumors revealed that the RT4 cells were surrounded by intact extracellular matrix, while the invasive EJ cells lacked extracellular matrix. 6 In our present study, we also demonstrated the effect on epithelial cell gene expression by an intact basement mem brane. When the invasive EJ cells were grown on plastic, because the cells were not capable of synthesizing the type IV collagen, the type IV procollagen message appeared to be up regulated (figure 3). However, when the invasive EJ cells were grown on an artificial basement membrane Matrigel, the type IV procollagen message was down-regulated (figure 4). The lack of significant hybridization to the type IV procollagen mRNA in the non-invasive RT4 cells can be explained by their ability to synthesize and deposit type IV collagen and laminin (figure lA). This observation is consistent with our hypothesis that the basement membrane may play a functional role in gene regulation. Others have shown that extracellular matrices have an affect, positively and dramatically, on the functional secretion of milk ·proteins in cultured mammary epithelial cells. 16 Streuli and BisseP7 have further shown that the levels of laminin, type IV collagen, and fibronectin mRNAs are highest in cells cultured on plastic surfaces and that the interaction between mammary
431
epithelial cells and both basement membrane and stromally derived matrices appears to exert a negative influence on the expression of mRNA for extracellular matrix components. We also found that the morphology of the EJ cell line also differed in the two environments. On plastic, the cells were spindle-shaped with a suggestion of pseudopodia-like projec tions (figure 2A). On Matrigel, the cells were clustered and had a papillary-like organization (figure 2B). One possible mechanism explaining these observations is based on known interconnections between the extracellular matrix, cytoskeleton, and nuclear matrix. 18 Coffey has proposed that the nuclear matrix actually forms an interlocking cellular matrix network by direct connection with the cytoskeletal matrix, and extends to the extracellular matrix. 18 The nuclear matrix is comprised of nuclear proteins called lamins, which only recently have been identified as members of the intermediate filament family.18 In addition to interac tions with chromosomes, the lamins also form a scaffold to which the nuclear pores are attached. Outside the nucleus, the cytoplasmic intermediate filaments are frequently close to the nuclear envelope. The cytoplasmic intermediate filaments may interact with the lamins, either directly via the nuclear pores or indirectly via a component of the nuclear envelope 18 (figure 5). The nuclear matrix is associated with RNA synthesis. Tran scriptional complexes have been identified on the nuclear ma trix. 19 O'Malley et al.20 observed that 95% of the unprocessed mRNA precursor for ovalbumin was associated with the nuclear matrix. When the intrans were processed out, the mature mRNA was released from the nuclear matrix. This suggests that the nuclear matrix was involved in mRNA processing and that any alterations in the nuclear matrix structures, which may occur as a result from the absence of an extracellular matrix, could alter important steps in transcription and RNA processing. Furthermore, Reid and her colleagues21 ' 22 demonstrated that the regulation of insulin messenger RNA abundance in cultured hepatocytes was dependent on the type of the extracellular matrix components on which the cells were grown. Therefore, the extracellular matrix can alter gene expression of cells in a specific manner. Not only are individual extracellular matrix components important but that the interaction of several of the components are involved in cell signaling and function. More recently, Fearon et al.23 identified a colorectal cancer gene which was deleted or altered in more than 70% of colorec tal cancers. This gene was found to specify a protein with sequences similar to those of the neural cell adhesion molecules. A gene altered in colorectal tumors which is similar in sequence to genes involved in cell-surface interactions is relevant to our present study of cell-basement membrane interactions. The extracellular matrix components contact transmembrane re ceptors through the cell membrane which includes the families of integrins and cell adhesion molecules. 24 These receptors contain sequences for phosphorylation in their cytoplasmic domain and may be important in providing signals that deter mine the location, polarity, and shape of cells. Signals from the extracellular matrix, therefore, may exert as much control over the behavior of cells as do hormones and other soluble mediators. Our results demonstrated that the basement membrane has a regulatory effect on both cellular morphology and nuclear function on urothelial cells. We also suggest that the absence of basement membrane associated with invasive human bladder tumor cells may not only facilitate their ability to penetrate into blood and lymphatic vessels, but may affect their gene expressions, contributing to further ge netic instability and leading to phenotypic variations which are the hallmark of malignancy. We are currently pursuing the molecular relationships between the extracellular matrix and the nuclear matrix. Results from our studies may elucidate the mechanisms responsible for tumor heterogeneity, as well as
432
SCHLUSSEL, DROLLER AND LIU
developmental anomalies such as Alport syndrome in which components of the extracellular matrix are also altered. Acknowledgments. The authors thank the assistance and friendship provided by Dr. Gregory Elder of the Center For Molecular Biology at Mt. Sinai School of Medicine. REFERENCES 1. Fidler, I. J. and Poste, G.: The cellular heterogeneity of malignant neoplasms: implications for adjuvant chemotherapy. Semin. On col., 12: 207, 1985. 2. Droller, M. J.: Transitional cell cancer: upper tracts and bladder. In: Campbell's Urology, ed. 5. Edited by P. C. Walsh, R. F. Gittes, A. D. Perlmutter, and T. A. Stamey. Vol. 2, pp. 13531366, Philadelphia: W.B. Saunders Co., 1986. 3. Nowell, P.: Mechanisms of tumor progression. Cancer Res., 46: 2203, 1986. 4. Conn, I. G., Crocker, J. and Wallace, D. M. A. et al.: Basement membranes in urothelial carcinoma. Br. J. Urol., 60: 536, 1987. 5. Liotta, L. A.: Tumor invasion and metastases: role of the basement membrane. Am. J. Pathol., 117: 339, 1984. 6. Weiss, R. E., Liu, B. C. S. and Ahlering, T. E. et al.: Mechanisms of human bladder tumor invasion: role of protease cathepsin B. J. Urol., 144: 798, 1990. 7. Redwood, S. M., Weiss, R. E., Droller, M. J. and Liu, B. C. S.: Abrogation of invasion by human bladder tumor cells using protease inhibitors. Surgical Forum, 41: 694, 1990. 8. Rigby, C. C. and Franks, L. M.: A human tissue culture cell line from a transitional cell tumor of the urinary bladder: growth, chromosome pattern, and ultrastructure. Br. J. Cancer, 24: 746, 1970. 9. Shih, C. and Weinberg, R. A.: Isolation of a transforming sequence from a human bladder carcinoma cell line. Cell, 29: 161, 1982. 10. Li, M. L., Aggeler, J. and Farson, D. A. et al.: Influence of a reconstituted basement membrane and its components on casein gene expression and secretion in mouse mammary epithelial cells. Proc. Natl. Acad. Sci. USA, 84: 136, 1987.
11. Chirgwin, J. M., Przybyla, A. L. and MacDonald, R. J. et al.: Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry, 18: 5294, 1979. 12. Maniatis, T. E., Fritsch, E. F. and Sambrook, J.: In: Molecular Cloning: A Laboratory Manual, p. 545. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982. 13. Fourney, R. M., Miyakoshi, J. and Day-III, R. S. et al.: Northern blotting: efficient RNA staining and transfer. Focus, 10: 5, 1987. 14. Feinberg, A. P. and Vogelstein, B. V.: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activ ity. Anal. Biochem., 137: 266, 1984. 15. Pihlajaniemi, T., Tyggvason, K. and Myers, J. C. et al.: cDNA clones for the pro-alpha 1 (IV) chain of human type IV procol lagen reveal an unusual homology of amino acid sequences in two halves of the carboxyl-terminal domain. J. Biol. Chem., 260: 7681, 1985. 16. Barcellos-Hoff, M. H., Aggeler, J. and Ram, T. G. et al.: Functional differentiation and alveolar morphogenesis of primary mammary epithelial cells cultured on reconstituted basement membrane. Development, 105: 223, 1989. 17. Streuli, C. H. and Bissell, M. J.: Expression of extracellular matrix components is regulated by substratum. J. Cell Biol., 110: 1405, 1990. 18. Nelson, W. G., Pienta, K. J., Barrack, E. R. and Coffey, D. S.: The role of the nuclear matrix in the organization and function of DNA. Ann. Rev. Biophys. Biophys. Chem., 15: 457, 1986. 19. Jackson, D. A., McCready, S. J. and Cook, P. R.: RNA is synthe sized at the nuclear cage. Nature, 292: 552, 1981. 20. Ciejek, E. M., Nordstromn, J. L., Tsai, M-J. and O'Malley, B. W.: Ribonucleic acid precursors are associated with the chick oviduct nuclear matrix. Biochemistry, 21: 4945, 1982. 21. Muschel, R., Khoury, G. and Reid, L. M.: Regulation of insulin mRNA abundance and adenylation: dependence on hormones and matrix substrate. Mo!. Cell Bio., 6: 337, 1986. 22. Reid, L. M.: Stem cell biology, hormone/matrix synergies and liver differentiation. Curr. Opin. Cell Biol., 2: 121, 1990. 23. Fearon, E. R., Cho, K. R. and Nigro, J. M. et al.: Identification of a chromosome 18q gene that is altered in colorectal cancers. Science, 247: 49, 1990. 24. Ruoslahti, E. and Pierschbacher, M. D.: New perspectives in cell adhesion: RGD and integrins. Science, 238: 491, 1987.
