- Email: [email protected]
The Mechanism of Human Bladder Tumor Implantation in an in Vitro Model
0022-534 7/86/1362-0482$02.00/0 Vol. 136, August Printed in U.S.A.
THE JOURNAL OF UROLOGY
Copyright © 1986 by The Williams & Wilkins Co.
THE MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION IN AN IN VITRO MODEL DOV PODE,* YAAKOV ALON, AVIVA T. HOROWITZ, ISRAEL VLODAVSKY AND SHOSHANA BIRAN From the Department of Urology, the Department of Cancer Research and the Department of Radiation and Clinical Oncology, Hebrew University-Hadassah Medical School, Jerusalem, Israel
ABSTRACT
Implantation of tumor cells in the bladder following transurethral resection of superficial bladder tumors is believed to be one factor in the etiology of bladder tumor recurrences. Using an in vitro model system we have studied the initial interaction between bladder carcinoma cells and a naturally produced basement membrane-like substrate. Minced explants of superficial low grade human bladder tumors from 10 patients were plated into culture dishes coated with a naturally produced extracellular matrix (ECM). This ECM has been shown to resemble the human urothelial basement membrane and submucosa in its macromolecular composition and ultrastructural appearance. It was found that a firm attachment of the human bladder tumor cells occurred within one hour, reached a maximal value within 24 hours and was followed by flattening and proliferation of the plated cells. These results indicate that prevention of tumor implantation should be initiated in the first hour after transurethral resection of the bladder tumors. This assay can be used for the investigation of various treatments to prevent tumor implantation. The main problem encountered in the treatment of noninvasive low grade bladder tumors is the high rate of recurrence, which ranges from 50 to 70 per cent, mostly within six to 12 months. 1 • 2 The failure of local treatment of superficial bladder cancer may be explained by the multicentricity of bladder tumors, incomplete excision of the primary tumor or continued exposure to carcinogens. Implantation of tumor cells during the initial TUR is another factor that may cause the high incidence of tumor recurrence. Clinical circumstantial evidence in humans and laboratory studies with animals indicate that this factor may be important. Recurrences at the urethra and bladder neck led a number of investigators to consider implantation as an etiologic factor in bladder tumor reappearance in humans. 3-6 Soloway and associates 7 •8 simulated the clinical situation by cauterization of a portion of the murine bladder before transurethral instillation of transitional tumor cells. Fifty-four to 94 per cent of previously cauterized bladders developed tumors in contrast to only 12 per cent when the cells were instilled into normal bladders. Shapiro et al. 9 using a similar model achieved an implantation and growth development rate of 69 per cent after cauterization of the bladder mucosa. In vitro studies performed by Vlodavsky et al. 10 demonstrated a much faster and more firm attachment of tumor cells to the subendothelial extracellular matrix as compared with an intact lining of endothelial cells. These experiments show that the neoplastic cells develop implants upon the denuded areas of the bladder mucosa mainly after surgery and not upon intact mucosa. A number of investigators have presumed that adjuvant intravesical chemotherapy may prevent implantation of tumor cells after TUR, without clearly understanding the mechanism of tumor cell adhesion to the extracellular matrix. Therefore most of the clinical studies have used intravesical chemotherapy starting one day to four weeks after resection of the tumors and performed for different periods of time. Several questions Accepted for publication February 27, 1986. * Requests for reprints: Dept. of Urology, Hebrew University, Hadassah Medical School, Jerusalem, Israel. Supported by the Joint Research Fund of the Hebrew University and Hadassah, and the Abraham and Irene Goldwater Memorial Fund.
remain to be answered before the best treatment can be designed. Should the first instillation be given early after surgery or delayeu to avoid systemic effects because of increased drug absorption? Should the intravesical chemotherapy be given in a single dose or prolonged treatment be advocated? An investigation into the mechanism of bladder tumor cell adhesion and its active flattening on exposed basement membrane or connective tissue is most pertinent to understanding the process of tumor cell implantation. The early stages of this phenomenon can be studied in vitro, by allowing human tumor cells to interact with a naturally produced basement membrane which closely resembles its in vivo counterpart. These requirements have been shown to be fulfilled by extracellular matrix (ECM) produced by cultured corneal endothelial cells. This ECM is similar to the bladder basement membrane in its appearance and characteristic components. 10 In the present study we report on the use of subendothelial ECM as an in vitro model substrate for bladder tumor implantation. MATERIALS AND METHODS
Tumors. Bladder tumors were obtained from 10 patients undergoing transurethral resection of superficial bladder tumors, using biopsy forceps prior to the electrical resection. Only low grade tumors were used in this study. Turners were minced to about one mm. pieces with a sharp scalpel knife in a sterile Petri dish and suspended in culture medium. Medium. Cells were maintained in serum free medium composed of 1:1 (v/v) mixture of Ham's F12 and Dulbecco's modified Eagle medium H-21 and supplemented with nonessential amino acids, vitamins, glutamine (Grand Island Biological Co.), antibiotics (40 µg./ml. gentamycin, 50 µ/ml. penicillin) and supplements as described by Biran et al. 11 Preparation of dishes coated with ECM. Bovine corneal endothelial cells were dissociated from steer eyes as described by Gospodarowicz et al., 12 plated at an initial density of 4 x 104 cells/35 mm. dish (Falcon) and maintained under the conditions described by Vlodavsky et al. 13 Six to eight days after reaching confluency the cultures were washed once with phos482
MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION
phate buffered saline (PBS) and exposed to 0.5 per cent Triton XlOO in PBS (30 min. of gentle shaking at room temperature). The cell layer was dissolved leaving the underlying ECM intact and firmly attached to the entire tissue culture dish. Analysis of the extracellular matrix by immunofluorescent and metabolic labelling revealed the presence of laminin, fibronectin, heparan sulfate and collagen types IV and III as the most abundant constituents. Adhesion assay. Finely minced tumor samples consisting of single cells and small aggregates consisting of 10 to 100 cells were suspended in medium and added (15 to 30 mg. tissue/2 ml.) to each of regular and ECM coated tissue culture dishes (Falcon). Cultures were maintained at 37C in a 10 per cent CO 2 humidified incubator. At various incubation times, the nonattached cells were removed by washes with PBS and the remaining firmly attached cells were dissociated into a single cell suspension with STV (0.05 per cent trypsin/0.02 per cent versene solution in PBS) and counted. Four dishes were scored for each determination. The variations in different determinations did not exceed ±15 per cent of the mean. Immunostaining of human bladder mucosa. Identification of adhesive glycoproteins in the urothelial basement membrane and submucosa was performed by indirect immunofluorescence staining of frozen sections of normal human bladder tissue. Biopsies from normal bladder mucosa were obtained from five patients during retropubic prostatectomy. All biopsies were immediately frozen in isopentan precooled in liquid nitrogen and sectioned in four slices. The slices were fixed in cold acetone for 20 minutes and incubated with either rabbit antilaminin, rabbit anticollagen type IV (at a dilution of 1:100) or rabbit antihuman plasma fibronectin, for 30 minutes in room temperature. The slices were then rinsed with PBS (three times for 15 minutes) and incubated (30 minutes, room temperature) with 1:20 fluorescein-conjugated goat anti-rabbit IgG (Welcome), and washed three times with PBS. Fluorescence was visualized by a Zeiss photomicroscope III equipped with phase contrast. Rabbit anti-mouse laminin and anti-collagen type IV antibodies were kindly provided by Dr. Victor Terranova, National Institute of Dental Research, Bethesda, Md. Rabbit antihuman plasma fibronectin was produced and characterized in our laboratory. Thymidine incorporation. To evaluate the extent of cell growth by thymidine incorporation the growth medium was replaced by one ml. of medium containing five µCi/ml. 3Hthymidine (16.2 Ci/mMol, New England Nuclear) and incubated for 48 hours. The medium was then removed and the cultures rinsed five times with PBS, fixed with metanol for 10
483
minutes and dried. The dishes were then coated with Ilford K5 photographic emulsion, incubated for five days, developed with Kodak D-19 developer and fixed in 10 per cent thiosulphate. RESULTS
Immunofluorescent staining performed on frozen sections of normal human bladder mucosa obtained from five patients during prostatectomy revealed that laminin fibronectin and collagen type IV are present in the urothelial basement membrane whereas the submucosa contains primarily fibronectin (fig. IA, B). Tumor cell implantation on ECM. Single tumor cells and aggregates suspended in serum free medium were seeded on ECM and tested for cell attachment, morphological appearance and 3H-thymidine incorporation. Cell attachment and flattening occurred within one hour after seeding as visualized by phase microscopy (fig. 2) and scanning electron microscopy (fig. 3). Single cells plated on ECM were highly flattened and showed broad peripheral lamellae which tapered to distal areas of discrete attachment to the ECM. In the case of cell aggregates, attachment and flattening was preceded by migration of cells out of the cell aggregates leading to the formation of small colonies composed of flat monolayer cells, within two to 24 hours (fig. 4). In contrast, cells seeded on regular plastic tissue culture dishes remained floating or loosely attached, mostly in the form of small aggregates composed of spheroidal cells. The extent and rate of cell attachment was evaluated by washing out the non-attached cells and counting remaining firmly attached cells following their dissociation into a single cell suspension with trypsin-EDTA. It was found that adhesion to ECM was far greater and faster than to regular culture dishes. Tumor cells plated on ECM reached maximal cell adherence within 24 hours (1.07 X 105 cells/mg. of tissue, range: 0.76 X 105 to 1.52 X 105 ). In contrast only few aggregates of tumor cells attached to plastic within 24 hours (4 x 102 cells/ mg. of tissue, range Oto 8 x 102) (fig. 5). Three to six days after seeding, cell piling and shedding into the culture medium was noticed. Hence, only 70 per cent of the cells remained attached to the ECM for three days and 35 per cent for seven days. Only viable cells exhibited cell attachment and flattening as demonstrated by the uptake of trypan blue. Eighty per cent of the cells which remained floating, 24 hours after seeding on ECM, were dead. In contrast, the attached and flattened cells on the ECM were 100 per cent viable. 3H-thymidine incorporation and autoradiography studies performed 24 hours after seeding the cells on ECM revealed a high labelling index as indicated by
FIG. 1. Immunofluorescent staining of laminin and fibronectin in human bladder mucosa. A, laminin in urothelial basement membrane and in blood vessels. B, fibronectin in urothelial basement and submucosa.
484
PODE AND ASSOCIATES
FIG. 2. Morphological appearance of human bladder tumor cells one hour after seeding on ECM, showing intensive attachment and flattening of single tumor cells and aggregates. Visualized by phase microscopy X200.
FIG. 4. Morphological appearance and culture organization of bladder tumor cells 24 hours after seeding on ECM. Visualized by phase contrast microscopy X200 showing flattened, closely apposed monolayer of tumor cells. 1.0
0.5
Plastic
FIG. 3. Scanning electronmicroscopy of human bladder tumor cell attached to ECM 2 hours after seeding.
3
7
Days in tissue culture
the appearance of labelled nuclei (fig. 6). These results demonstrate that cell attachment to the ECM was followed by DNA synthesis and active proliferation. The various constituents of the culture medium in this assay had only a minor effect on the attachment of tumor cells during the first 24 hours. The use of PBS or cystoscopy fluid (1.1 per cent glycine in water) instead of culture medium did not significantly change the rate and extent of tumor cell adhesion. The morphological appearance as well as attachment rate of tumor cells from the various tissue samples were uniform. DISCUSSION
The "implantation theory" of tumors first proposed by Albarran and Imbert in 190314 postulates that exfoliated tumor cells can reimplant directly on the urothelium. Although implantation of cancer cells upon intact mucosa has been discounted by most pathologists, the possibility of seeding on raw surfaces or in surgical wounds is widely accepted. 3• 6 Tumor cell attachment to the basement membrane or the extracellular matrix is a necessary prerequisite for the development of a recurrent superficial tumor. Various observations indicate that the basement membrane and ECM in the submucosa do not function merely as an inert structural support but rather play
FIG. 5. Extent of human bladder tumor cell attachment to ECM or plastic as function of time in culture. Tumor cells were seeded into regular or ECM coated dishes. Tumor cell attachment to ECM was at maximum within 24 hours.
an active role in the control of cellular adhesion, flattening, migration, proliferation and differentiation of cells. 13• 15 The low adhesiveness of the intact epithelium is due to its unidirectional secretion of the extracellular material under the basal cell layer exclusively. The apical surface lacks the adhesive proteins and is therefore nonadherent. 10 These findings correlate well with the clinical observations on the preferential localization of bladder tumor recurrences in areas with a defective urothelium. The nature and specifity of the adhesive components in the ECM have been studied extensively by a number of groups using different models. 15- 17 The extracellular matrix contains collagenous and noncollagenous glycoproteins and glycoaminoglycans. Collagens, fibronectin and laminin are the major adhesive proteins in the ECM. Biochemically, the process of cell adhesion consists of molecular interactions between various cell surface constituents, receptor sites and adhesive glycoproteins, such as laminin, fibronectin and various types of collagen in the ECM. Direct contact of epithelial cells with the ECM
MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION
485
apy immediately after the tumor resection aimed at preventing tumor cell reimplantation. Burnard et al. 22 instilled 90 mg. thiotepa into the bladder immediately after tumor resection. There was a significant diminution of the recurrences at one year. Similar results were reported by Zincke and associates. 23 They concluded that additional treatment later during the follow-up interval did not further improve results. In view of the rapid attachment of bladder cells to the exposed basement membrane and ECM we suggest that every prophylactic intravesical chemotherapy, either one dose or multiple dose protocols, should be initiated within the first hour postoperatively.
REFERENCES
FIG. 6. Thymidine incorporation pattern of bladder tumor cells cultured on ECM coated culture dishes. 24 hours after seeding, cultures were subjected to thymidine labelling and autoradiography. Labelled nuclei were distributed throughout cell layer as visualized by phase microscopy X200.
and their subsequent flattening are thought to be required for the appropriate organization of their cytoskeleton and for the subsequent stimulation of the cytoplasmic machinery. 17 When the extent of cell spreading was correlated with DNA synthesis or cell growth it was found to be highly coupled. 18 In the present study we have identified the presence of laminin fibronectin and collagen type IV in the urothelial basement membrane. Fibronectin was found in the submucosa. These components are present in the ECM produced by bovine corneal endothelial cells in vivo and in vitro. ECM coated dishes, therefore provided a model with which to investigate the attachment of bladder tumor cells to the urothelial basement membrane as well as agents capable of blocking this attachment. Although such an in-vitro model might not correlate quantitatively with the human in vivo tumor cell implantation process, our results show that adherence of human bladder tumor cells plated in direct contact with an exposed basement membrane occurred during the first hour and reached a maximum within 24 hours. After the first 24 hours we observed a continuous decline in the number of attached cells due to cell piling and shedding into the culture medium. Although our model might not predict the in vivo rate of tumor cell proliferation and subsequent development of new tumors, some conclusions should be made. Since the attachment of epithelial cells to ECM exerts a permissive effect on cell proliferation, 18 prophylactic treatments should be aimed at destroying floating tumor cells or preventing their adhesion at the time of surgery or immediately afterwards. Tumor cells which did not attach to the basement membrane or submucosa during the first 24 hours most probably will not adhere later. Minimizing surgical trauma and thorough irrigation of the bladder with the cystoscopy fluids may not eliminate tumor cell implantation after TUR. Therefore, the use of cytotoxic agents as an irrigant has been suggested. Various prophylactic intravesical chemotherapy protocols have used very different schedules. The time to initiate intravesical chemotherapy has been chosen arbitrarily. 19- 21 Only a few clinical studies have used single dose chemother-
1. Greene, L. F., Hanash, K. A. and Farrow, G. M.: Benign papilloma or papillary carcinoma of the bladder. J. Urol., 110: 205, 1973. 2. Frankson, C.: Tumors of the urinary bladder: a pathological and clinical solution of 434 cases. Acta Chir. Scand. (Suppl.), 151: 186, 1950. 3. Boreham, P.: The surgical spread of cancer in urology. Br. J. Urol., 28: 163, 1956. 4. Weyranch, H. M. and Crossfield, J. H.: Dissemination of bladder neoplasms by endoscopic electrocoagulation. J. Urol., 87: 391, 1962. 5. Kiefer, J. H.: Bladder tumor recurrence in the urethra; a warning. J. Urol., 69: 652, 1953. 6. Hinman, F. Jr.: Recurrence of bladder tumors by surgical implantation. J. Urol., 75: 695, 1956. 7. Soloway, M. S. and Masters, S.: Urothelial susceptability to tumor cell implantation influence of cauterization. Cancer, 46: 1158, 1980. 8. Soloway, M. S., Nissenkorn, I., McCallum, L. W. and Murphy, W. M.: Single and sequential combination intravesical chemotherapy of murine bladder cancer. Urology, 19: 169, 1982. 9. Shapiro, A., Kelley, D. R., Oakley, D. M., Catalona, W. J. and Ratliff, T. L.: Technical factors affecting the reproducibility of intravesical mouse bladder tumor implantation during therapy with Bacillus Calmette Guerin. Cancer Res., 44: 3051, 1984. 10. Vlodavsky, I., Arian, Y., Atzmon, R. and Fuks, Z.: Tumor cell attachment to the vascular endothelium and subsequent degredation of the subendothelial extracellular matrix. Exp. Cell Res., 140: 149, 1982. 11. Biran, S., Horowitz, A. T., Fuks, Z. and Vlodavsky, I.: High density lipoprotein and extracellular matrix promotes growth and plating efficiency of normal human mammary epithelial cells in serum free medium. Int. J. Cancer, 31: 557, 1983. 12. Gospodarowicz, D., Mescher, A. R. and Birdwell, C.R.: Stimulation of corneal endothelial cell proliferation in vitro by fibroblast and epidermal growth factors. Exp. Eye Res., 25: 75, 1977. 13. Vlodavsky, I., Lui, G. M. and Gospodarowicz, D.: Morphological appearance, growth behaviors and migratory activity of human cells maintained on extracellular matrix versus plastic. Cell, 19: 607, 1980. 14. Albarran, J. and Imbert, L.: Les tumeurs du rein. Paris, Masson et Cie, pp. 452-459, 1903. 15. Kleinman, H. K., Klebe, R. J. and Martin, G. R.: Role of collagenous matrices in the adhesion and growth of cells. J. Cell Biol., 88: 473, 1981. 16. Alitalo, K. and Vaheri, A.: Pericellular matrix in malignant transformation. Adv. Cancer Res., 37: 111, 1983. 17. Hay, E. D.: Extracellular matrix. J. Cell Biol., 91: 205, 1981. 18. Folkman, J. and Moscona, A.: Role of cell shape in growth control. Nature, 273: 345, 1978. 19. Soloway, M. S.: Rational for intensive intravesical chemotherapy for superficial bladder cancer. J. Urol., 123: 461, 1980. 20. Koontz, W. W., Prout, G. R. Jr., Smith, W., Frable, W. J. and Minnis, J. E.: The use of intravesical thiotepa in the management of noninvasive carcinoma of the bladder. J. Urol., 125: 307, 1981. 21. Schulman, C.: Prophylactic chemotherapy of superficial transitional cell bladder carcinoma, an EORTC randomized trial comparing thiotepa, epipodophyllotoxin (VM26) and TUR alone. Eur. Urol., 8: 207, 1982. 22. Burnard, K. G., Boyd, P. J. R., Mays, M. E., Shuttleworth, K. E.
486
PODE AND ASSOCIATES
D. and Lloyd-Davies, R. W.: Single dose intravesical thiotepa as an adjuvant to cystodiathermy in the treatment of transitional cell bladder carcinoma. Br. J. Urol., 48: 55, 1976. 23. Zincke, H., Utz, D. C., Taylor, W. F., Myers, R. P. and Leary, F.
J.: Influence of thiotepa and doxorubicin instillation at the time of transurethral surgical treatment of bladder cancer on tumor recurrence: prospective randomized, double blind controlled trial. J. Urol., 129: 505, 1983.
THE JOURNAL OF UROLOGY
Copyright © 1986 by The Williams & Wilkins Co.
THE MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION IN AN IN VITRO MODEL DOV PODE,* YAAKOV ALON, AVIVA T. HOROWITZ, ISRAEL VLODAVSKY AND SHOSHANA BIRAN From the Department of Urology, the Department of Cancer Research and the Department of Radiation and Clinical Oncology, Hebrew University-Hadassah Medical School, Jerusalem, Israel
ABSTRACT
Implantation of tumor cells in the bladder following transurethral resection of superficial bladder tumors is believed to be one factor in the etiology of bladder tumor recurrences. Using an in vitro model system we have studied the initial interaction between bladder carcinoma cells and a naturally produced basement membrane-like substrate. Minced explants of superficial low grade human bladder tumors from 10 patients were plated into culture dishes coated with a naturally produced extracellular matrix (ECM). This ECM has been shown to resemble the human urothelial basement membrane and submucosa in its macromolecular composition and ultrastructural appearance. It was found that a firm attachment of the human bladder tumor cells occurred within one hour, reached a maximal value within 24 hours and was followed by flattening and proliferation of the plated cells. These results indicate that prevention of tumor implantation should be initiated in the first hour after transurethral resection of the bladder tumors. This assay can be used for the investigation of various treatments to prevent tumor implantation. The main problem encountered in the treatment of noninvasive low grade bladder tumors is the high rate of recurrence, which ranges from 50 to 70 per cent, mostly within six to 12 months. 1 • 2 The failure of local treatment of superficial bladder cancer may be explained by the multicentricity of bladder tumors, incomplete excision of the primary tumor or continued exposure to carcinogens. Implantation of tumor cells during the initial TUR is another factor that may cause the high incidence of tumor recurrence. Clinical circumstantial evidence in humans and laboratory studies with animals indicate that this factor may be important. Recurrences at the urethra and bladder neck led a number of investigators to consider implantation as an etiologic factor in bladder tumor reappearance in humans. 3-6 Soloway and associates 7 •8 simulated the clinical situation by cauterization of a portion of the murine bladder before transurethral instillation of transitional tumor cells. Fifty-four to 94 per cent of previously cauterized bladders developed tumors in contrast to only 12 per cent when the cells were instilled into normal bladders. Shapiro et al. 9 using a similar model achieved an implantation and growth development rate of 69 per cent after cauterization of the bladder mucosa. In vitro studies performed by Vlodavsky et al. 10 demonstrated a much faster and more firm attachment of tumor cells to the subendothelial extracellular matrix as compared with an intact lining of endothelial cells. These experiments show that the neoplastic cells develop implants upon the denuded areas of the bladder mucosa mainly after surgery and not upon intact mucosa. A number of investigators have presumed that adjuvant intravesical chemotherapy may prevent implantation of tumor cells after TUR, without clearly understanding the mechanism of tumor cell adhesion to the extracellular matrix. Therefore most of the clinical studies have used intravesical chemotherapy starting one day to four weeks after resection of the tumors and performed for different periods of time. Several questions Accepted for publication February 27, 1986. * Requests for reprints: Dept. of Urology, Hebrew University, Hadassah Medical School, Jerusalem, Israel. Supported by the Joint Research Fund of the Hebrew University and Hadassah, and the Abraham and Irene Goldwater Memorial Fund.
remain to be answered before the best treatment can be designed. Should the first instillation be given early after surgery or delayeu to avoid systemic effects because of increased drug absorption? Should the intravesical chemotherapy be given in a single dose or prolonged treatment be advocated? An investigation into the mechanism of bladder tumor cell adhesion and its active flattening on exposed basement membrane or connective tissue is most pertinent to understanding the process of tumor cell implantation. The early stages of this phenomenon can be studied in vitro, by allowing human tumor cells to interact with a naturally produced basement membrane which closely resembles its in vivo counterpart. These requirements have been shown to be fulfilled by extracellular matrix (ECM) produced by cultured corneal endothelial cells. This ECM is similar to the bladder basement membrane in its appearance and characteristic components. 10 In the present study we report on the use of subendothelial ECM as an in vitro model substrate for bladder tumor implantation. MATERIALS AND METHODS
Tumors. Bladder tumors were obtained from 10 patients undergoing transurethral resection of superficial bladder tumors, using biopsy forceps prior to the electrical resection. Only low grade tumors were used in this study. Turners were minced to about one mm. pieces with a sharp scalpel knife in a sterile Petri dish and suspended in culture medium. Medium. Cells were maintained in serum free medium composed of 1:1 (v/v) mixture of Ham's F12 and Dulbecco's modified Eagle medium H-21 and supplemented with nonessential amino acids, vitamins, glutamine (Grand Island Biological Co.), antibiotics (40 µg./ml. gentamycin, 50 µ/ml. penicillin) and supplements as described by Biran et al. 11 Preparation of dishes coated with ECM. Bovine corneal endothelial cells were dissociated from steer eyes as described by Gospodarowicz et al., 12 plated at an initial density of 4 x 104 cells/35 mm. dish (Falcon) and maintained under the conditions described by Vlodavsky et al. 13 Six to eight days after reaching confluency the cultures were washed once with phos482
MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION
phate buffered saline (PBS) and exposed to 0.5 per cent Triton XlOO in PBS (30 min. of gentle shaking at room temperature). The cell layer was dissolved leaving the underlying ECM intact and firmly attached to the entire tissue culture dish. Analysis of the extracellular matrix by immunofluorescent and metabolic labelling revealed the presence of laminin, fibronectin, heparan sulfate and collagen types IV and III as the most abundant constituents. Adhesion assay. Finely minced tumor samples consisting of single cells and small aggregates consisting of 10 to 100 cells were suspended in medium and added (15 to 30 mg. tissue/2 ml.) to each of regular and ECM coated tissue culture dishes (Falcon). Cultures were maintained at 37C in a 10 per cent CO 2 humidified incubator. At various incubation times, the nonattached cells were removed by washes with PBS and the remaining firmly attached cells were dissociated into a single cell suspension with STV (0.05 per cent trypsin/0.02 per cent versene solution in PBS) and counted. Four dishes were scored for each determination. The variations in different determinations did not exceed ±15 per cent of the mean. Immunostaining of human bladder mucosa. Identification of adhesive glycoproteins in the urothelial basement membrane and submucosa was performed by indirect immunofluorescence staining of frozen sections of normal human bladder tissue. Biopsies from normal bladder mucosa were obtained from five patients during retropubic prostatectomy. All biopsies were immediately frozen in isopentan precooled in liquid nitrogen and sectioned in four slices. The slices were fixed in cold acetone for 20 minutes and incubated with either rabbit antilaminin, rabbit anticollagen type IV (at a dilution of 1:100) or rabbit antihuman plasma fibronectin, for 30 minutes in room temperature. The slices were then rinsed with PBS (three times for 15 minutes) and incubated (30 minutes, room temperature) with 1:20 fluorescein-conjugated goat anti-rabbit IgG (Welcome), and washed three times with PBS. Fluorescence was visualized by a Zeiss photomicroscope III equipped with phase contrast. Rabbit anti-mouse laminin and anti-collagen type IV antibodies were kindly provided by Dr. Victor Terranova, National Institute of Dental Research, Bethesda, Md. Rabbit antihuman plasma fibronectin was produced and characterized in our laboratory. Thymidine incorporation. To evaluate the extent of cell growth by thymidine incorporation the growth medium was replaced by one ml. of medium containing five µCi/ml. 3Hthymidine (16.2 Ci/mMol, New England Nuclear) and incubated for 48 hours. The medium was then removed and the cultures rinsed five times with PBS, fixed with metanol for 10
483
minutes and dried. The dishes were then coated with Ilford K5 photographic emulsion, incubated for five days, developed with Kodak D-19 developer and fixed in 10 per cent thiosulphate. RESULTS
Immunofluorescent staining performed on frozen sections of normal human bladder mucosa obtained from five patients during prostatectomy revealed that laminin fibronectin and collagen type IV are present in the urothelial basement membrane whereas the submucosa contains primarily fibronectin (fig. IA, B). Tumor cell implantation on ECM. Single tumor cells and aggregates suspended in serum free medium were seeded on ECM and tested for cell attachment, morphological appearance and 3H-thymidine incorporation. Cell attachment and flattening occurred within one hour after seeding as visualized by phase microscopy (fig. 2) and scanning electron microscopy (fig. 3). Single cells plated on ECM were highly flattened and showed broad peripheral lamellae which tapered to distal areas of discrete attachment to the ECM. In the case of cell aggregates, attachment and flattening was preceded by migration of cells out of the cell aggregates leading to the formation of small colonies composed of flat monolayer cells, within two to 24 hours (fig. 4). In contrast, cells seeded on regular plastic tissue culture dishes remained floating or loosely attached, mostly in the form of small aggregates composed of spheroidal cells. The extent and rate of cell attachment was evaluated by washing out the non-attached cells and counting remaining firmly attached cells following their dissociation into a single cell suspension with trypsin-EDTA. It was found that adhesion to ECM was far greater and faster than to regular culture dishes. Tumor cells plated on ECM reached maximal cell adherence within 24 hours (1.07 X 105 cells/mg. of tissue, range: 0.76 X 105 to 1.52 X 105 ). In contrast only few aggregates of tumor cells attached to plastic within 24 hours (4 x 102 cells/ mg. of tissue, range Oto 8 x 102) (fig. 5). Three to six days after seeding, cell piling and shedding into the culture medium was noticed. Hence, only 70 per cent of the cells remained attached to the ECM for three days and 35 per cent for seven days. Only viable cells exhibited cell attachment and flattening as demonstrated by the uptake of trypan blue. Eighty per cent of the cells which remained floating, 24 hours after seeding on ECM, were dead. In contrast, the attached and flattened cells on the ECM were 100 per cent viable. 3H-thymidine incorporation and autoradiography studies performed 24 hours after seeding the cells on ECM revealed a high labelling index as indicated by
FIG. 1. Immunofluorescent staining of laminin and fibronectin in human bladder mucosa. A, laminin in urothelial basement membrane and in blood vessels. B, fibronectin in urothelial basement and submucosa.
484
PODE AND ASSOCIATES
FIG. 2. Morphological appearance of human bladder tumor cells one hour after seeding on ECM, showing intensive attachment and flattening of single tumor cells and aggregates. Visualized by phase microscopy X200.
FIG. 4. Morphological appearance and culture organization of bladder tumor cells 24 hours after seeding on ECM. Visualized by phase contrast microscopy X200 showing flattened, closely apposed monolayer of tumor cells. 1.0
0.5
Plastic
FIG. 3. Scanning electronmicroscopy of human bladder tumor cell attached to ECM 2 hours after seeding.
3
7
Days in tissue culture
the appearance of labelled nuclei (fig. 6). These results demonstrate that cell attachment to the ECM was followed by DNA synthesis and active proliferation. The various constituents of the culture medium in this assay had only a minor effect on the attachment of tumor cells during the first 24 hours. The use of PBS or cystoscopy fluid (1.1 per cent glycine in water) instead of culture medium did not significantly change the rate and extent of tumor cell adhesion. The morphological appearance as well as attachment rate of tumor cells from the various tissue samples were uniform. DISCUSSION
The "implantation theory" of tumors first proposed by Albarran and Imbert in 190314 postulates that exfoliated tumor cells can reimplant directly on the urothelium. Although implantation of cancer cells upon intact mucosa has been discounted by most pathologists, the possibility of seeding on raw surfaces or in surgical wounds is widely accepted. 3• 6 Tumor cell attachment to the basement membrane or the extracellular matrix is a necessary prerequisite for the development of a recurrent superficial tumor. Various observations indicate that the basement membrane and ECM in the submucosa do not function merely as an inert structural support but rather play
FIG. 5. Extent of human bladder tumor cell attachment to ECM or plastic as function of time in culture. Tumor cells were seeded into regular or ECM coated dishes. Tumor cell attachment to ECM was at maximum within 24 hours.
an active role in the control of cellular adhesion, flattening, migration, proliferation and differentiation of cells. 13• 15 The low adhesiveness of the intact epithelium is due to its unidirectional secretion of the extracellular material under the basal cell layer exclusively. The apical surface lacks the adhesive proteins and is therefore nonadherent. 10 These findings correlate well with the clinical observations on the preferential localization of bladder tumor recurrences in areas with a defective urothelium. The nature and specifity of the adhesive components in the ECM have been studied extensively by a number of groups using different models. 15- 17 The extracellular matrix contains collagenous and noncollagenous glycoproteins and glycoaminoglycans. Collagens, fibronectin and laminin are the major adhesive proteins in the ECM. Biochemically, the process of cell adhesion consists of molecular interactions between various cell surface constituents, receptor sites and adhesive glycoproteins, such as laminin, fibronectin and various types of collagen in the ECM. Direct contact of epithelial cells with the ECM
MECHANISM OF HUMAN BLADDER TUMOR IMPLANTATION
485
apy immediately after the tumor resection aimed at preventing tumor cell reimplantation. Burnard et al. 22 instilled 90 mg. thiotepa into the bladder immediately after tumor resection. There was a significant diminution of the recurrences at one year. Similar results were reported by Zincke and associates. 23 They concluded that additional treatment later during the follow-up interval did not further improve results. In view of the rapid attachment of bladder cells to the exposed basement membrane and ECM we suggest that every prophylactic intravesical chemotherapy, either one dose or multiple dose protocols, should be initiated within the first hour postoperatively.
REFERENCES
FIG. 6. Thymidine incorporation pattern of bladder tumor cells cultured on ECM coated culture dishes. 24 hours after seeding, cultures were subjected to thymidine labelling and autoradiography. Labelled nuclei were distributed throughout cell layer as visualized by phase microscopy X200.
and their subsequent flattening are thought to be required for the appropriate organization of their cytoskeleton and for the subsequent stimulation of the cytoplasmic machinery. 17 When the extent of cell spreading was correlated with DNA synthesis or cell growth it was found to be highly coupled. 18 In the present study we have identified the presence of laminin fibronectin and collagen type IV in the urothelial basement membrane. Fibronectin was found in the submucosa. These components are present in the ECM produced by bovine corneal endothelial cells in vivo and in vitro. ECM coated dishes, therefore provided a model with which to investigate the attachment of bladder tumor cells to the urothelial basement membrane as well as agents capable of blocking this attachment. Although such an in-vitro model might not correlate quantitatively with the human in vivo tumor cell implantation process, our results show that adherence of human bladder tumor cells plated in direct contact with an exposed basement membrane occurred during the first hour and reached a maximum within 24 hours. After the first 24 hours we observed a continuous decline in the number of attached cells due to cell piling and shedding into the culture medium. Although our model might not predict the in vivo rate of tumor cell proliferation and subsequent development of new tumors, some conclusions should be made. Since the attachment of epithelial cells to ECM exerts a permissive effect on cell proliferation, 18 prophylactic treatments should be aimed at destroying floating tumor cells or preventing their adhesion at the time of surgery or immediately afterwards. Tumor cells which did not attach to the basement membrane or submucosa during the first 24 hours most probably will not adhere later. Minimizing surgical trauma and thorough irrigation of the bladder with the cystoscopy fluids may not eliminate tumor cell implantation after TUR. Therefore, the use of cytotoxic agents as an irrigant has been suggested. Various prophylactic intravesical chemotherapy protocols have used very different schedules. The time to initiate intravesical chemotherapy has been chosen arbitrarily. 19- 21 Only a few clinical studies have used single dose chemother-
1. Greene, L. F., Hanash, K. A. and Farrow, G. M.: Benign papilloma or papillary carcinoma of the bladder. J. Urol., 110: 205, 1973. 2. Frankson, C.: Tumors of the urinary bladder: a pathological and clinical solution of 434 cases. Acta Chir. Scand. (Suppl.), 151: 186, 1950. 3. Boreham, P.: The surgical spread of cancer in urology. Br. J. Urol., 28: 163, 1956. 4. Weyranch, H. M. and Crossfield, J. H.: Dissemination of bladder neoplasms by endoscopic electrocoagulation. J. Urol., 87: 391, 1962. 5. Kiefer, J. H.: Bladder tumor recurrence in the urethra; a warning. J. Urol., 69: 652, 1953. 6. Hinman, F. Jr.: Recurrence of bladder tumors by surgical implantation. J. Urol., 75: 695, 1956. 7. Soloway, M. S. and Masters, S.: Urothelial susceptability to tumor cell implantation influence of cauterization. Cancer, 46: 1158, 1980. 8. Soloway, M. S., Nissenkorn, I., McCallum, L. W. and Murphy, W. M.: Single and sequential combination intravesical chemotherapy of murine bladder cancer. Urology, 19: 169, 1982. 9. Shapiro, A., Kelley, D. R., Oakley, D. M., Catalona, W. J. and Ratliff, T. L.: Technical factors affecting the reproducibility of intravesical mouse bladder tumor implantation during therapy with Bacillus Calmette Guerin. Cancer Res., 44: 3051, 1984. 10. Vlodavsky, I., Arian, Y., Atzmon, R. and Fuks, Z.: Tumor cell attachment to the vascular endothelium and subsequent degredation of the subendothelial extracellular matrix. Exp. Cell Res., 140: 149, 1982. 11. Biran, S., Horowitz, A. T., Fuks, Z. and Vlodavsky, I.: High density lipoprotein and extracellular matrix promotes growth and plating efficiency of normal human mammary epithelial cells in serum free medium. Int. J. Cancer, 31: 557, 1983. 12. Gospodarowicz, D., Mescher, A. R. and Birdwell, C.R.: Stimulation of corneal endothelial cell proliferation in vitro by fibroblast and epidermal growth factors. Exp. Eye Res., 25: 75, 1977. 13. Vlodavsky, I., Lui, G. M. and Gospodarowicz, D.: Morphological appearance, growth behaviors and migratory activity of human cells maintained on extracellular matrix versus plastic. Cell, 19: 607, 1980. 14. Albarran, J. and Imbert, L.: Les tumeurs du rein. Paris, Masson et Cie, pp. 452-459, 1903. 15. Kleinman, H. K., Klebe, R. J. and Martin, G. R.: Role of collagenous matrices in the adhesion and growth of cells. J. Cell Biol., 88: 473, 1981. 16. Alitalo, K. and Vaheri, A.: Pericellular matrix in malignant transformation. Adv. Cancer Res., 37: 111, 1983. 17. Hay, E. D.: Extracellular matrix. J. Cell Biol., 91: 205, 1981. 18. Folkman, J. and Moscona, A.: Role of cell shape in growth control. Nature, 273: 345, 1978. 19. Soloway, M. S.: Rational for intensive intravesical chemotherapy for superficial bladder cancer. J. Urol., 123: 461, 1980. 20. Koontz, W. W., Prout, G. R. Jr., Smith, W., Frable, W. J. and Minnis, J. E.: The use of intravesical thiotepa in the management of noninvasive carcinoma of the bladder. J. Urol., 125: 307, 1981. 21. Schulman, C.: Prophylactic chemotherapy of superficial transitional cell bladder carcinoma, an EORTC randomized trial comparing thiotepa, epipodophyllotoxin (VM26) and TUR alone. Eur. Urol., 8: 207, 1982. 22. Burnard, K. G., Boyd, P. J. R., Mays, M. E., Shuttleworth, K. E.
486
PODE AND ASSOCIATES
D. and Lloyd-Davies, R. W.: Single dose intravesical thiotepa as an adjuvant to cystodiathermy in the treatment of transitional cell bladder carcinoma. Br. J. Urol., 48: 55, 1976. 23. Zincke, H., Utz, D. C., Taylor, W. F., Myers, R. P. and Leary, F.
J.: Influence of thiotepa and doxorubicin instillation at the time of transurethral surgical treatment of bladder cancer on tumor recurrence: prospective randomized, double blind controlled trial. J. Urol., 129: 505, 1983.