D 2000 Nature America, Inc. 0929-1903/00/$15.00/+0 www.nature.com/cgt
An improved intravesical model using human bladder cancer cell lines to optimize gene and other therapies Takafumi Watanabe,1 Nobuo Shinohara,2 Ataru Sazawa,2,3 Toru Harabayashi,2 Yoshifumi Ogiso,4 Tomohiko Koyanagi,2 Mitsuyoshi Takiguchi,1 Akira Hashimoto,1 Noboru Kuzumaki,3 Motoyuki Yamashita,5 Motoyoshi Tanaka,6 H. Barton Grossman,6 and William F. Benedict5,6 1
Department of Veterinary Clinical Sciences, Graduate School of Veterinary Medicine, Hokkaido University, Sapporo, Japan; 2Department of Urology, Cancer Institute, Hokkaido University School of Medicine, Sapporo, Japan; 3Division of Gene Regulation, Cancer Institute, Hokkaido University School of Medicine, Sapporo, Japan; 4Department of Pathology, Nagano Children's Hospital, Nagano, Japan; 5Department of Genitourinary Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030; and 6Department of Urology, The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030.
Orthotopic implantation of human bladder cancer cells into immunodeficient mice is an important tool for studying the biology and effects of therapy. Nevertheless, the incidence of tumor implantation and growth by transurethral instillation of the human bladder cancer cells into murine bladders has been low or not reproducible. However, using a modified intravesical technique and the human bladder cancer cell lines, KU - 7 and UM - UC - 2, we have been able to obtain a high and reproducible incidence of superficial bladder tumors. Furthermore, intravesical administration of the LacZ adenovirus vector resulted in significant - galactosidase expression in these bladder tumors as well as the normal urothelium, which was associated with the removal of the glycosoaminoglycan layer. Because this modified technique produces a high incidence of superficial human tumor growth and allows the efficacy of gene transfer to be evaluated, it should be a useful model for the study of intravesical gene therapy for human bladder cancer. Cancer Gene Therapy ( 2000 ) 7, 1575 ± 1580 Key words: Orthotopic intravesical model; human bladder cancer; gene therapy; adenoviral vector.
I
mplantation of bladder cancer cells into mice has been reported to be a useful procedure for the evaluation of various therapeutic modalities. Orthotopic animal models of bladder cancer more closely mimic the behavior of these tumors in humans than nonorthotopic (usually subcutaneous ) models. Initially, orthotopic growth was achieved by inoculating syngeneic bladder cancer cells into immunocompetent mice 1 ± 4 or implanting the human bladder cancer cells xenographically into immunodeficient mice.5 ± 8 The latter approach is particularly important for the characterization of human bladder cancer biology including patterns of growth and differentiation, expression of tumor antigens, and response to the treatment. Although the basic technique of orthotopic implantation of human bladder cancer cells by transurethral instillation has been used extensively by various investigators, tumor incidences
Received June 12, 2000; accepted September 5, 2000. Address correspondence and reprint requests to Dr. Nobuo Shinohara, Department of Urology, Hokkaido University School of Medicine, Kita - 15, Nishi - 7 Kita - ku, Sapporo, 060 - 8638 Japan. E-mail address: nobuo -
[email protected]
Cancer Gene Therapy, Vol 7, No 12, 2000: pp 1575 ± 1580
range from 6% to 50% and were not reproducible, particularly following intravesical instillation of the tumor.5 ± 8 Therefore, we determined that a more reliable method was needed for establishing intravesical human bladder tumors and wished to further develop such a system for intravesical therapy. Local recurrence of superficial bladder cancer despite aggressive local therapy is common, and the prognosis of advanced bladder cancer with intensive multimodal therapy is poor.9,10 New therapeutic strategies, including gene therapy, are being evaluated to improve the outcome of patients with this disease, and a variety of gene therapy strategies are being considered for the next generation of human cancer therapy. Adenoviral vectors have many advantages as the means to delivering gene therapy compared with retroviral and other vectors, especially for intravesical gene therapy.11,12 To aid in the development of optimal intravesical model for gene and other therapies, we have developed a modified technique for the establishment of reproducible intravesical human bladder cancers in athymic mice. Furthermore, we demonstrate significant adenoviral 1575
1576
WATANABE, SHINOHARA, SAZAWA, ET AL: HUMAN BLADDER CANCER CELL MODEL
mediated gene transfer throughout the full thickness of normal urothelium and into the superficial bladder tumors by intravesical instillation of an adenoviral vector containing a LacZ gene insert, but only after removal of the glycosoaminoglycan (gag) layer. MATERIALS AND METHODS Bladder cancer cell lines
Two human bladder cancer cell lines, KU -7 (provided by Dr. M. Tachibana, Keio University, Japan ) and UM - UC - 2, were used.13,14 Both cell lines were maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum and 0.05% glutamine in a 378C incubator with a humidified environment of 5% CO2 in air. A total of 293 embryonic kidney cells were obtained from the American Type Culture Collection and grown in the same culture medium as above. Orthotopic administration of human bladder cancer cells
The intravesical orthotopic animal model using human bladder cancer cells was a modification of the technique described by Ahlering et al.5 Briefly, female KSN athymic nude mice (Japan SLC, Shizuoka, Japan ) were anesthetized with an intraperitoneal injection of a 2.5% solution of 1:1 ( vol /vol) 2,2,2 - tribromoethanol ( Sigma, St. Louis, MO ) and tert -amyl alcohol (Wako, Osaka, Japan ). Either a 22 - or 24- gauge catheter was inserted into the bladder transurethrally and 100 L of 0.2% trypsin in 0.02% EDTA was infused, which was retained in the bladder for 30 minutes. After trypsinization, the bladder was recatheterized, washed with phosphate -buffered saline ( PBS ), and scratched carefully by the catheter. Subsequently, a 100 -L suspension of PBS containing 1107 KU -7 cells or UM UC -2 cells was instilled into the bladder. The urethra was ligated with 6- 0 nylon suture to assure that the cells were retained in the bladder ( Fig 1) . After 3 hours the suture was removed, and the bladder was evacuated by spontaneous voiding. Incidence of bladder cancer
Ten days following the intravesical administration of the bladder cancer cells, the mice were sacrificed to confirm the tumor incidence. Bladders were distended with 250 L of 10% buffered formalin, and removed. Following fixation, the bladders were paraffin embedded, serial sectioned, and stained with hematoxylin and eosin. Twenty sections were then taken from each mouse and the incidence of tumors was established microscopically. Gene transfer of adenoviral vectors expressing LacZ
The replication - defective adenoviral vector AdCMV-LacZ ( provided by N. Senmaru, Hokkaido University School of Medicine, Sapporo, Japan ) contains an expression cassette
Figure 1. The schematic ventral view of the mouse into which the human bladder cancer cell lines are instilled. After trypsinization and scratching the bladder mucosa, a suspension of 1107 human bladder cancer cells was instilled into the bladder. Urination was stopped for 3 hours by ligating the urethra with a purse - string suture.
encoding the Escherichia coli - galactosidase ( LacZ ) gene under transcriptional regulation of the immediate early cytomegalovirus (CMV ) promoter and upstream of the SV40 polyadenylation signal.15 Viral vectors were expanded following repeated infection of 293 cells in spinner culture. - Galactosidase expression in vitro
A total of 5104 cells per well were plated into six -well plates. The following day, the cells were infected with AdCMV-LacZ at MOIs ranging from 10 to 100. Forty -eight hours after infection, the cells were fixed with 1.25% glutaraldehyde for 5 minutes and stained with 0.6 mg /ml of X -gal in 5 Mm potassium ferricyanide, 5 Mm potassium ferrocyanide, and 2 Mm MgCl2 in PBS at 378C, overnight. Infection efficiency was calculated by the following ratio: positive cells/total number of cells100%. - Galactosidase expression in vivo
Mice were anesthetized and recatheterized, and 100 L of a suspension of AdCMV-LacZ (1109 pfu /100 L per day) was administered for 1 hour on consecutive days, ranging from 1 to 3 days, i.e., one to three total instillations. The urethra was ligated similarly to the procedure used to inject bladder cancer cells described above to insure a consistent 1 hour exposure to the adenovirus. Forty -eight hours after administration of AdCMV-LacZ, the mice were sacrificed. For analysis of - galactosidase expression in vivo, the bladders were distended with OCT- compound ( Sakura Finetechnical, Tokyo, Japan ) , resected, and frozen in liquid nitrogen. Frozen sections were fixed in 1.25% glutaraldehyde for 5 minutes, incubated with X - gal, (0.6 mg /ml ) , overnight at 378C in a humidified atmosphere and counterstained with eosin. Cancer Gene Therapy, Vol 7, No 12, 2000
WATANABE, SHINOHARA, SAZAWA, ET AL: HUMAN BLADDER CANCER CELL MODEL
RESULTS Orthotopic intravesical administration of human bladder cancer cell lines
KU -7 and UM -UC - 2 cells were instilled intravesically into the mouse bladder. Ten days after administration of human bladder cancer cells, the mice were sacrificed and the bladder was removed. Tumors were confirmed microscopically. KU 7 tumors exhibited a growth pattern that usually protruded into the bladder ( Fig 2A and B ), whereas UM -UC -2 cells tended to grow initially in a pattern resembling carcinoma in situ with spreading beneath the normal urothelium (Fig 2C and D ). For both cell lines the experiments were performed three times ( Table 1 ). When the KU -7 cells were used
1577
procedure -related deaths occurred in 17 ±30% of the mice and produced tumors in 100% of the surviving mice. In experiments using UM -UC -2 cells, procedure -related death occurred in 0 ±30% of the mice with tumors being formed in 80% of the surviving mice. Adenoviral transgene expression in the orthotopic bladder cancer model
To demonstrate adenoviral transgene expression in this orthotopic bladder cancer model, we examined the expression of -galactosidase after infection with the recombinant adenovirus, AdCMV-LacZ. Initially the infectivity and transgene expression of the adenovirus in KU - 7 and UM UC -2 cells was examined in vitro. To infect > 90% of the cells, an MOI of 25 (KU -7 ) or 10 ( UM -UC -2 ) was
Figure 2. Microscopic appearance of superficial KU - 7 ( A and B ) or UM - UC - 2 ( C and D ) tumors following hematoxylin ± eosin staining of paraffin - embedded bladder sections 10 days after intravesical instillation. Magnifications shown are 10 ( A and C ) or 40 ( B and D ) . The arrows in ( A ) and ( B ) show the same area at high and low powers. The broken arrow in ( C ) shows the border between the normal urothelium and the tumor growth.
Cancer Gene Therapy, Vol 7, No 12, 2000
1578
WATANABE, SHINOHARA, SAZAWA, ET AL: HUMAN BLADDER CANCER CELL MODEL
Table 1. Human Bladder Cancer Orthotopic Model In Nude Mice
Bladder cancer cell line KU - 7
UM - UC - 2
Experiment number
Number of mice used
Number of procedure related deaths* ( mortality rate, % )
1 2 3 total 1 2 3 total
10 9 12 31 5 6 7 18
3 2 2 7 ( 23 ) 1 2 0 3 ( 17 )
Number of mice that survived
Number of tumors ( incidence, % ) y
7 7 10 24 4 4 7 15
7 7 10 24 ( 100 ) 4 4 4 12 ( 80 )
* Deaths were scored as procedure related up to 5 days after inoculation of tumor cells into the bladder. y Ten days after intravesical instillation of the cancer cells. Tumors were confirmed microscopically.
Figure 3. Examples of - galactosidase transgene expression in vivo. - Galactosidase expression is shown by the blue stain that included in certain areas the entire thickness of the epithelium and additional cells in the lamina propria ( A ) as well as parts of a small KU - 7 tumor shown with an arrow. The bladder was removed 6 days after the KU - 7 cells were instilled and following two daily intravesical exposures of AdCMV - LacZ ( 1109 pfu / day ) for 1 hour. A matching paraffin - embedded section is shown after hematoxylin ± eosin staining with the arrow again showing the small KU - 7 tumor. Note the metaplasia seen as a result of the trypsin treatment. Shown in ( C ) is the only area in the entire bladder that had slight - galactosidase expression following three similar daily intravesical treatments where no previous exposure to trypsin or tumor injection had occurred. This small amount of staining in fact could have resulted from mechanical manipulation of the catheter, which removed a portion of the gag layer. ( D ) shows an adjacent paraffin - embedded section of the same bladder seen in ( C ) . Note the lack of significant metaplasia in the absence of trypsin treatment.
Cancer Gene Therapy, Vol 7, No 12, 2000
WATANABE, SHINOHARA, SAZAWA, ET AL: HUMAN BLADDER CANCER CELL MODEL
required. However, at lower MOIs the infectivity was also high, with little evidence of viral produced toxicity. Based on these in vitro results, -galactosidase expression in the orthotopic bladder cancer model was examined following intravesical administration of AdCMV- LacZ. The frequency of transduction was first examined in nine mice in whom KU - 7 cells had been instilled. These were assigned to three groups. In group I (n =3 ), AdCMV- LacZ ( 1109 pfu /day ) was administered three times for 1 hour at 2, 3, and 4 days after tumor inoculation. In group II ( n = 3) and group III ( n= 3 ), AdCMV- LacZ was administered twice at 3 and 4 days or once at 4 days in the same manner, respectively. All mice were sacrificed 6 days after tumor inoculation and the expression of - galactosidase was examined. Microscopically, little -galactosidase expression was observed in the bladders of mice that received one treatment with AdCMV-LacZ (group III ). In groups I and II, the expression of - galactosidase was seen not only in the tumor cells but also through the full thickness of the normal urothelium and into some normal cells within the lamina propria (Fig 3A ) . In UM - UC -2 cells infected two or three times with AdCMV- LacZ in a similar manner and time frame as KU - 7 cells, comparable - galactosidase expression was seen in tumor cells and normal urothelium. To investigate the effect of adenoviral transgene expression in the normal bladder, we performed transurethral administration of AdCMV-LacZ into the normal mouse bladder three times for 1 hour on consecutive days without trypsin pretreatment or tumor inoculation. In contrast to the above results in which trypsin was instilled before the AdCMV-LacZ administration, - galactosidase expression was not observed to any significant extent (Fig 3C ). DISCUSSION
For many studies, orthotopic bladder cancer models provide a more relevant environment to investigate the biology of and response to therapy than subcutaneous models. One orthotopic approach consists of surgical direct injection of tumor cells into the bladder wall after exposing the bladder by an abdominal incision.16 ± 18 Although this technique can result in almost 100% tumor take for specific cell lines, it requires an open surgical procedure and does not reflect the behavior of superficial bladder cancer. Nor does the tumor usually extend superficially through the normal urothelium, and thus, cannot be used to optimize intravesical therapy for superficial bladder cancer. Another orthotopic model has used transurethral administration of bladder cancer cells into the bladder by retrograde catheterization that mimics superficial bladder cancer.1 ± 8 However, this technique has a low and nonreproducible tumor incidence. To increase the incidence of tumor formation, traumatization of the bladder mucosa has been employed using such agents as N- methyl -N -nitro-
Cancer Gene Therapy, Vol 7, No 12, 2000
1579
sourea or hydrochloric acid,7 as well as scraping the bladder wall with a catheter.1 ± 7 High tumor incidence has been reported in syngeneic models of murine bladder cancer using these procedures.2 ± 4 Gunther et al 2 for example found 100% tumor incidence in a syngeneic murine mouse model after traumatizing the bladder mucosa with a catheter using the MB49 murine bladder cancer cell line. Nevertheless, a human bladder intravesical model that is reproducible has been unavailable to date. In the present study, we used trypsin and mechanical injury with the catheter to traumatize the bladder mucosa as a pretreatment before tumor inoculation. The retention time was also modified by using a purse -string suture to assure that the cancer cells remained in the bladder for 3 hours. This modified procedure provided a high tumor incidence ( 80± 100% ) of human superficial bladder tumors and required no major surgical intervention. Procedure -related deaths in the first 5 days after tumor inoculation were seen at the beginning of our experience. We speculate that these procedure -related failures were associated with anesthesia or catheterization -related injuries with bladder perforation. In more recent studies using the same methodology, procedure -related deaths have been minimal. Gene therapy using transurethral administration of an adenoviral vector is a promising novel therapy for bladder cancer. We examined adenoviral transgene expression in our orthotopic bladder cancer model. - Galactosidase expression using AdCMV- LacZ was seen not only throughout the superficial tumors, but also throughout the normal urothelium that extended into some cells within the lamina propria. Werthman et al4 also demonstrated some adenoviral transgene expression in a syngeneic murine orthotopic bladder cancer model. We believe, however, that this report is the first to document adenoviral transgene expression in a human bladder cancer orthotopic model by intravesical administration. This finding suggests that this model can be used for the preclinical evaluation of human cancer gene therapy using adenoviral and other vectors. It is important to note, however, that - galactosidase was not uniformly expressed throughout the bladders infected with AdCMV-LacZ. Moreover, the removal of the gag layer appeared necessary for effective intravesical adenoviral gene therapy in the bladder, because minimal gene transfer was seen if the bladder had not been exposed to trypsin within 2± 3 days before intravesical adenoviral instillation. Such treatment with trypsin at the concentration used has previously been shown to produce rapid extensive urothelial shedding and presumably loss of the associated gag layer.19 Recently, Engler et al20 reported that pretreatment of the bladder with ethanol was effective for improving adenoviral -mediated gene transfer and expression in the bladder urothelium, consistent with our results indicating the importance of the initial removal of the gag layer before intravesical gene therapy. Further studies in our laboratory will focus on methods to remove the gag layer using reagents
1580
WATANABE, SHINOHARA, SAZAWA, ET AL: HUMAN BLADDER CANCER CELL MODEL
that can be used in human intravesical trials, because the results to date strongly suggest that such a strategy will be required to obtain a high level of gene transfer in superficial bladder tumors by intravesical instillation. ACKNOWLEDGMENT
This work was supported by a Grant -in -Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture, Japan, from the Retina Research Foundation and from Tobacco Settlement Funds as Appropriated by the Texas State Legislature. REFERENCES 1. Bonfil RD, Russo DM, Schmilovich AJ, et al. Intravesical therapy with vinorelbine tartrate: antitumor activity in orthotopic murine cell carcinoma of the bladder. J Urol. 1997;158:912 ± 915. 2. Gunther JH, Jurczok A, Wulf T, et al. Optimizing syngeneic orthotopic murine bladder cancer ( MB49 ) . Cancer Res. 1999;59:2834 ± 2837. 3. Lee KE, Weiss GH, O'Donnell RW, et al. Reduction of bladder cancer growth in mice treated with intravesical Bacillus Calmette Guerin and systemic interleukin 2. J Urol. 1987;137:1270 ± 1273. 4. Werthman PE, Drazan KE, Rosenthal JT, et al. Adenoviral - p53 gene transfer to orthotopic and peritoneal murine bladder cancer. J Urol. 1996;155:753 ± 756. 5. Ahlering TE, Dubeau L, Jones PA. A new in vivo model to study invasion and metastasis of human bladder carcinoma. Cancer Res. 1987;47:6660 ± 6665. 6. Soloway MS, Masters S. Urothelial susceptibility to tumor cell implantation: influence of cauterization. Cancer. 1980;46:1158 ± 1163. 7. Soloway MS, Nissenkorn I, McCallum L. Urothelial susceptibility to tumor cell implantation: comparison of cauterization with N - methyl N - nitrosourea. Urology. 1983;21:159 ± 161.
8. Theodorescu D, Cornil I, Fernandez BJ, et al. Overexpression of normal and mutated forms of HRAS induces orthotopic bladder invasion in a human transitional cell carcinoma. Proc Natl Acad Sci USA. 1990;87:9047 ± 9051. 9. Corral DA, Logothetis CJ. New approaches in the treatment of metastatic transitional - cell cancer of the bladder. World J Urol. 1997; 15:139 ± 143. 10. Warde P, Gospodarowicz MK. New approaches in the use of radiation therapy in the treatment of infiltrative transitional - cell cancer of the bladder. World J Urol. 1997;15:125 ± 133. 11. Graham FL, Prevec L. Methods for construction of adenovirus vectors. Mol Biotech. 1995;3:207 ± 220. 12. Mufson MA, Belshe RB. A review of adenoviruses in the etiology of acute hemorrhagic cystitis. J Urol. 1976;115:191 ± 194. 13. Tachibana M, Miyakawa A, Uchida A, et al. Granulocyte colony stimulating factor receptor expression on human transitional cell carcinoma of the bladder. Br J Cancer. 1997;75:1489 ± 1496. 14. Grossman HB, Wedemeyer G, Ren L. UM - UC - 1 and UM - UC - 2: characterization of two new human transitional cell carcinoma lines. J Urol. 1984;132:834 ± 837. 15. Senmaru N, Shichinohe T, Takeuchi M, et al. Suppression of Erk activation and in vivo growth in esophageal cancer cells by the dominant negative Ras mutant, N116Y. Int J Cancer. 1998;78:366 ± 371. 16. Chang SG, Kim JI, Jung JC, et al. Antimetastatic activity of the new platinum analog [ Pt ( cis - dach ) ( DPPE ) .2NO3 ] in a metastatic model of human bladder cancer. Anticancer Res. 1997;17:3239 ± 3242. 17. Jiang F, Zhou XM. A model of orthotopic murine bladder ( MBT- 2 ) tumor implants. Urol Res. 1997;25:179 ± 182. 18. Kleinerman DI, Dinney CP, Zhang WW, et al. Suppression of human bladder cancer growth by increased expression of C - CAM1 gene in an orthotopic model. Cancer Res. 1996;56:3431 ± 3435. 19. Gardiner,RA, Masters JRW, Molland EA. Experimental trypsin induced urothelial separation in vivo. Br J Urol. 1980;52:295 ± 300. 20. Engler H, Anderson SC, Machemer TR, et al. Ethanol improves adenovirus - mediated gene transfer and expression to the bladder epithelium of rodents. Urology. 1999;53:1049 ± 1053.
Cancer Gene Therapy, Vol 7, No 12, 2000