- Email: [email protected]
Integrin αv and coxsackie adenovirus receptor expression in clinical bladder cancer
BASIC SCIENCE
INTEGRIN ␣v AND COXSACKIE ADENOVIRUS RECEPTOR EXPRESSION IN CLINICAL BLADDER CANCER MARKUS D. SACHS, KATHERINE A. RAUEN, MEERA RAMAMURTHY, JENNIFER L. DODSON, ANGELO M. DE MARZO, MATHEW J. PUTZI, MARK P. SCHOENBERG, AND RONALD RODRIGUEZ
ABSTRACT Objectives. To evaluate the expression of the coxsackie and adenovirus receptor (CAR) and ␣v integrins in clinical specimens of bladder cancer to determine the susceptibility to adenoviral gene therapy. Efficient adenovirus-based gene therapy requires binding of the virus to CAR and involves the ␣v integrins. Studies on bladder cancer cell lines have shown that low adenoviral transduction rates were associated with low-level expression of CAR. Integrin ␣v expression increases in various tumors suggest its importance in differentiation, proliferation, and migration. CAR is structurally a member of the Ig-type superfamily of cell-cell adhesion molecules, suggesting that its expression may also be related to the state of tumor differentiation. Methods. We performed immunohistochemistry for CAR and integrin ␣v expression in bladder cancer specimens in 50 paraffin-embedded tumor-normal pairs and confirmed the results by quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of 11 separate bladder tumors and 4 separate normal bladder controls. Results. Immunochemistry demonstrated a stage and grade-dependent decrease in CAR expression (90.0%, 83.3%, and 31.3% of normal urothelium and superficial and invasive transitional cell carcinoma [TCC] and 83.3% and 39.5% of low and high-grade TCC, respectively). Furthermore, we found a stage and grade-dependent increase in ␣v integrin expression (13.3%, 46.0%, and 56.3% of normal urothelium, superficial TCC, and invasive TCC and 25% and 52.6% of low and high-grade TCC, respectively). Quantitative RT-PCR analysis confirmed a downregulation at the CAR gene expression level. Conclusions. This down-regulation may have a major impact on developing adenoviral-based gene therapy modalities. In addition, we propose that loss of CAR expression decreases rigid cell adhesion, possibly increasing the migratory potential. Loss of CAR expression correlates with the invasive phenotype in our analysis of bladder cancer. Simultaneously, the finding of increased ␣v expression in invasive cancer suggests a pathogenesis that involves heterophilic adhesion and migration of these cells on various extracellular ligands. UROLOGY 60: 531–536, 2002. © 2002, Elsevier Science Inc.
C
arcinoma of the bladder is the second most common cancer of the urogenital tract and makes up approximately 6% of all malignancies. In This work was supported in part by a research grant to M. D. Sachs from the Max Kade Foundation, New York, New York. From the Brady Urological Institute, Johns Hopkins Hospital, Baltimore, Maryland; Department of Urology, Charite Medical School, Humboldt University, Berlin, Germany; Department of Pediatrics, University of California, San Francisco, School of Medicine; Cancer Research Institute, Comprehensive Cancer Center, San Francisco, California; and Johns Hopkins Oncology Center, Department of Surgical Pathology, Johns Hopkins Hospital, Baltimore, Maryland Reprint requests: Ronald Rodriguez, M.D., Ph.D., Brady Urological Institute, Marburg 205, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287 Submitted: December 7, 2001, accepted (with revisions): April 1, 2002 © 2002, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED
2001, an estimated 54,300 people will have been diagnosed and an estimated 12,400 patients will have died of bladder cancer in the United States.1 Bladder cancer treatment failure is a morbid condition with limited options. Gene therapy directed against these failures represents a novel and evolving alternative.2 Adenovirus binding and uptake are separate but cooperative events. Binding requires interaction between the knob domain of the adenoviral fiber protein present at each vertex of the icosahedral structure of the virus and the coxsackie and adenovirus receptor (CAR),3 and internalization involves a secondary interaction between the penton base and certain patterns of integrins, in particular, ␣v3 and ␣v5.4 CAR is a 46-kD, class I, transmembrane protein with two Ig-like C2 domains. Struc0090-4295/02/$22.00 PII S0090-4295(02)01748-X 531
turally it is a member of the Ig superfamily of adhesion molecules and under certain conditions has been shown to suppress cell growth in vitro.5,6 CAR has a structure highly homologous to other known adhesion molecules, such as A33 antigen7 and vascular endothelial junction-associated protein.8 Integrins are cell-surface adhesion molecules that regulate normal cellular interactions, such as cell migration and assembly of, and interaction with, extracellular matrix ligands (eg, vitronectin) by binding Arg-Gly-Asp (RGD) amino acid sequences,9 cell growth and differentiation, and cell survival. These heterodimeric integral membrane proteins are composed of an ␣-chain and a -chain and facilitate signal transduction in a bidirectional manner, thereby having a direct impact on cell behavior.10 –12 Aberrant integrin expression is believed to play a role in tumor invasion and metastasis. Integrin ␣v plays a variety of roles in carcinogenesis, including differentiation, proliferation, prevention of apoptosis, matrix degradation, and adhesion.13–16 De novo ␣v expression in esophageal squamous cell carcinoma has been suggested to decrease rigid homophilic cell adhesion, possibly increasing the migratory potential while simultaneously permitting the heterophilic adhesion and migration of malignant cells on various ligands.17 Integrins ␣v3 and ␣v5 play a role in glioma-associated angiogenesis, neoangiogenesis, and cell migration in high-grade gliomas.18 Wechsel et al.19 described a grade-dependent increase in ␣v3 expression in renal cell carcinoma as a step in metastasis formation by increasing adhesion to the basement membrane. Although adenoviral binding requires an interaction with CAR, internalization is at least partly mediated by an interaction between the viral penton base and the ␣v3 or ␣v5 integrins. Takayama et al.20 found a correlation between the level of ␣v5 integrin and the expression level of a transferred gene and suggested that quantification of ␣v5 expression may be a good way of predicting the susceptibility of cells to adenoviral infection. In bladder cancer cell lines, the binding of adenovirus by way of CAR seems to be impaired. Li et al.21 reported that several human bladder cancer cell lines lacked expression of CAR, although the genomic sequences for the gene were intact. This downregulation correlated with resistance of urothelial cells to adenoviral infection.21 More recently, Okegawa et al.6 showed an apparent stagedependent reduction of CAR mRNA in clinical bladder cancer specimens using a semiquantitative polymerase chain reaction (PCR) method. A reduction in the actual CAR protein expression in bladder cancer, if proved, would have a major impact on the development of new adenoviral-based gene therapies. We present a detailed analysis of CAR and inte532
grin expression performed on bladder cancer specimens. In addition to predicting the clinical susceptibility to adenoviral gene transfer, these data may also provide valuable insight into the development of the invasive phenotype in bladder cancers. MATERIAL AND METHODS CELL LINE HEK293 cells (human embryo kidney, a packaging cell line for adenovirus, known to express very high levels of CAR) were purchased from American Tissue Culture Collection (Manassas, Va). Cells were grown in Dulbecco’s modification of Eagle’s medium (Mediatech), supplemented with 10% fetal bovine serum and 0.1% gentamicin.
TISSUE SAMPLES Normal and cancerous paraffin-embedded formalin-fixed bladder specimens from 50 patients (40 men and 10 women) were obtained from the pathology archives of Johns Hopkins Hospital for immunohistochemical analysis. Tumor samples were available from all 50 patients; normal urothelium was only available from 30 patients. Permission to evaluate these tissues was obtained from our institutional review board. Tissue sections were arranged in a tissue microarray, as previously described, using 0.6-mm cores.22 The mean patient age was 63.7 ⫾ 9.5 years (range 41 to 77). Tumor stage was assigned according to the American Joint Committee on Cancer staging 1997 TNM system. The pathologic stage was as follows: pTa, n ⫽ 7; pTis, n ⫽ 5; pT1, n ⫽ 6; pT2, n ⫽ 6; pT3, n ⫽ 16; and pT4, n ⫽ 10. Snap-frozen additional specimens of normal urothelium (n ⫽ 4) and bladder carcinoma Stage Tis (n ⫽ 1), Ta (n ⫽ 6), T1 (n ⫽ 2), T2 (n ⫽ 1), and T3 (n ⫽ 1) were obtained for quantitative reverse transcriptase (RT)-PCR analysis.
IMMUNOHISTOCHEMISTRY FOR CAR AND ␣V Integrin Expression Tissue sections were deparaffinized in xylene and dehydrated in ethanol. For CAR staining, standard indirect immunoperoxidase methods were used for immunostaining with rabbit polyclonal CAR 72 (Onyx Pharmaceuticals, Richmond, Calif), as previously described.23 In brief, the slides were baked for 30 minutes at 60°C and standard antigen retrieval methods, including trypsinization and microwave treatment in 10 mM citrate buffer, were performed. The tissue was blocked in 10% goat serum/phosphate-buffered saline and incubated in primary polyclonal antibody CAR 72 diluted 1:7000 for 8 to 16 hours at 4°C. The tissue sections were incubated in secondary biotinylated goat anti-rabbit immunoglobulin diluted 1:200 (Vector Laboratories, Burlingame, Calif) and then treated with streptavidin-biotinylated horseradish peroxidase complex (Vectastain Elite ABC Kit, Vector Laboratories). The sections were subsequently developed using diaminobenzidine tetrahydrochloride (Sigma) in hydrogen peroxide/phosphate-buffered saline and counterstained with hematoxylin. Staining for integrin ␣v was done using a monoclonal antibody (P2W7, Santa Cruz Biotechnology, Santa Cruz, Calif) at a 1:50 dilution with an overnight incubation and the LSAB2 System-HRP (Dako, Carpinteria, Calif). Both stainings were characterized in five staining levels (0, no staining; 1, 1% to 25% of cells showing membrane bound expression; 2, 25% to 50%; 3, 50% to 75%; and 4, 75% to 100%) by a single uropathologist (A.M.D.). For the purposes of presentation, the staining result was considered positive if more than 50% of cells expressed the antigen (staining levels 3 and 4) and negative if 50% or less of the cells showed a staining signal (levels 0 to 2). The staining is shown in Figure 1. UROLOGY 60 (3), 2002
denaturing for 15 seconds at 94°C, annealing for 30 seconds at 59.6°C and extending for 45 seconds at 70°C for 38 cycles, followed by a 7-minute final extension at 72°C.
STATISTICAL ANALYSIS The immunohistochemistry results were analyzed in relation to tumor stage and grade using the Kruskal-Wallis test for the average staining level and the chi-square test for the staining results (positive for staining levels 0 to 2 and negative for staining levels 3 to 4). Staining of tumor versus normal tissue was analyzed using the chi-square test. A result was considered statistically significant when the P value was less than 0.05. Statistical analysis was done using StatView for Windows, version 5.0.
RESULTS
FIGURE 1. CAR and integrin ␣v immunohistochemistry. (A) CAR-positive normal urothelium. (B) CAR-positive noninvasive transitional cell carcinoma. (C) CAR-negative invasive transitional cell carcinoma. (D) ␣v-negative normal urothelium. (E) ␣v-negative noninvasive transitional cell carcinoma. (F) ␣v-positive invasive transitional cell carcinoma (original magnification ⫻ 400).
REAL-TIME QUANTITATIVE RT-PCR FOR CAR RNA isolates were prepared from HEK293 cells and from fresh frozen tissue (normal bladder controls, n ⫽ 4, and bladder tumors, n ⫽ 11), using MicroPoly(A) Pure Kit (Ambion, Austin, Tex). PCRs were set up in a reaction volume of 50 L containing 23 L of reaction mix (Life Technologies, Rockville, Md), 8 L of 5 mM magnesium sulfate (Life Technologies), 10 L RNA, 2 L of each 5 M amplification primer, 1 L of 1:25,000 Sybr Green (Roche, Indianapolis, Ind), and 1 L of rt-Taq (Life Technologies). The sense and antisense primer sequence was 5⬘-TTGCTTGCTCTAGCGCTCATTGGTC-3⬘ and 5⬘-TCATCACAGGAATCGCACCCATTCG-3⬘, respectively (product size 325 bp). The amounts of target DNA (relative CAR copy number per haploid genome) were quantified using the following equation: Relative CAR copy number per haploid genome ⫽
CAR copy number . -actin copy number
The CAR copy number was calculated by generating a standard curve based on the amplification of known amounts of a plasmid carrying the CAR sequence. RNA values were then normalized by amplifying the constitutively expressed housekeeping gene -actin. The -actin copy number was calculated using serial dilutions of a known amount of -actin mRNA. Reactions were carried out in the iCycler thermal cycler (Bio-Rad, Hercules, Calif). The thermal cycling protocol was 30 minutes at 50°C and 2 minutes at 94°C, with subsequent UROLOGY 60 (3), 2002
INTEGRIN ␣V AND CAR PROTEIN EXPRESSION A summary of the staining results and staining levels is given in Figures 1 and 2. Staining levels 1 through 5 were assigned according to the percentage of cells that showed membranous staining (see Material and Methods section). To assess the staining results, we defined tumors as positive if more than 50% of cells showed a membranous staining signal (staining levels 3 and 4). Because this is a subjective definition, the actual staining level is also shown in a graph (Fig. 2C,D). Using these criteria, 50.0% (25 of 50) of the bladder tumors examined were considered negative for CAR expression compared with 10% (3 of 30) of the normal urothelium (obtained from the same bladder specimen). This downregulation was dependent on the tumor stage and grade, with 83.3% (15 of 18) versus 31.3% (10 of 32) expression in superficial (Ta, T1, Tis) and invasive (T2-T4) tumors, respectively, and 83.3% (10 of 12) versus 39.5% (15 of 38) in well and poorly differentiated tumors, respectively (Fig. 2A,B). The results were statistically significant when assessing the average staining level (0 to 4) and tumor stage (P ⫽ 0.0014, Kruskal-Wallis test), staining result (positive or negative) and tumor stage (P ⫽ 0.013, chi-square test), average staining level and tumor grade (P ⫽ 0.028, Kruskal-Wallis test), and staining result and tumor grade (P ⫽ 0.049, chi-square test). CAR expression in normal urothelium was equal in umbrella cells and more basal layers. CAR staining was positive in 90.0% of normal urothelium compared with in 50% of tumors, which was statistically significant (P ⫽ 0.0005, chi-square test). No significant differences were found between normal urothelial components from samples of the various stages and grades of bladder cancer specimens. For ␣v integrin, we found an overall positive staining of 46.0% (23 of 50) in bladder tumors and 13.3% (4 of 30) of normal urothelium. This difference was statistically significant (P ⫽ 0.003, chisquare test). The increase in ␣v expression was stage and grade dependent, with 56.3% (18 of 32) 533
FIGURE 2. CAR and integrin ␣v expression in bladder tumors. (A) Percentage of tumors with CAR staining level 3 or 4 (more than 50% of cells positive) and tumor stage and grade. (B) Average CAR staining level and tumor stage and grade. (C) Percentage of tumors with integrin ␣v staining level 3 or 4 and tumor stage and grade. (D) Average integrin ␣v staining level and tumor stage and grade.
of invasive tumor samples versus 27.8% (5 of 18) of superficial tumor samples and 52.6% (20 of 38) of high-grade tumor samples versus 25% (3 of 12) of low-grade tumor samples (Figure 2C,D) showing an increase. However, these values did not reach statistical significance (P ⫽ 0.234, chi-square test for comparison with stage, and P ⫽ 0.246, chi-square test for comparison with grade). The integrin ␣v staining in normal urothelium was more prominent in umbrella cells than in more basal layers. Stromal cells were universally negative for both CAR and integrin ␣v. REAL-TIME PCR Because the CAR antibody is an unvalidated, noncommercial reagent, we confirmed the CAR immunohistochemistry findings with real-time quantitative RT-PCR in a separate subset of samples obtained prospectively. As found in the immunohistochemistry results, a marked down-regulation of CAR occurred in bladder tumors (n ⫽ 11) compared with normal urothelium (n ⫽ 4), supporting the finding that CAR expression is downregulated in invasive bladder cancer. The HEK293 cells served as a positive control (Fig. 3). 534
COMMENT Because CAR and ␣v integrins are important for adenoviral binding and internalization, we sought to evaluate their expression patterns in the same samples of clinical bladder cancer. Interestingly, we found that the more invasive cancers tended to have an increased ␣v integrin pattern and a decreased CAR pattern of expression. The CAR expression data were more dramatic and easily reached statistical significance for this relatively small sample size. Although the overall difference of ␣v integrin expression between tumor and normal tissue was statistically significant, no significant difference was noted when segregated by stage and grade. Nonetheless, there clearly appeared to be a trend toward increased ␣v expression with stage and grade, which requires larger sample numbers to better assess. Recent analysis of bladder cancer cell lines has demonstrated marked down-regulation of CAR expression, as well as a significant resistance to transgene expression by adenoviral vectors.21 In addition, preclinical models of bladder cancer gene therapy have demonstrated an unusually high requirement for adenoviral vectors.24 This raised the UROLOGY 60 (3), 2002
FIGURE 3. Evaluation of CAR expression by quantitative RT-PCR. Average copy number with standard error.
question of whether adenoviral-based gene transduction into human bladder cancer cells would be possible. Previously, CAR expression in clinical bladder cancer samples had only been analyzed semiquantitatively by mRNA expression, suggesting an apparent down-regulation in the invasive tumors.6 We present a larger, more detailed analysis based on the protein expression pattern. Moreover, without an immunohistologic correlation, it is also unknown whether this down-regulation applies to all urothelial cells or to subpopulations. A reduction in active receptors would clearly lead to decreased viral gene delivery. Because CAR seems to play a role in cell-cell adhesion and migration, such a reduction might cause cells to lose cohesiveness and therefore may be a major contributing factor in the process of developing the invasive phenotype. We demonstrated by immunohistochemistry that clinical specimens of invasive bladder cancers reveal a marked stage and grade-dependent downregulation of CAR. Carcinoma in situ was evaluated as a superficial cancer, although one should bear in mind that this is a poorly differentiated tumor with high invasive potential. However, if carcinoma in situ is taken out of the data analysis, neither the overall rate (46.7%) nor the CAR positive rate of superficial tumors (84.6%) changed dramatically. Although the number of available samples was still low, these data suggest that CAR expression varies with the invasive potential of the cancer cells in clinical specimens of bladder cancer. We independently confirmed the marked downregulation of CAR in bladder tumors compared with normal urothelium by real-time quantitative RT-PCR. UROLOGY 60 (3), 2002
Our findings have two major clinical impacts. First, by characterizing the adenoviral receptor status as a function of clinical stage and grade, rational targeting of the appropriate patient populations suitable for gene therapy trials can be predicted. Moreover, for those patients who lack the appropriate receptor status, either modification to the adenoviral vectors will be required to allow high efficiency gene transfer or CAR expression will need to be reconstituted. Gene transfer with nonadenovirus vectors might be an alternative approach in patients with low CAR-expressing tumors. Second, these results provide important information about the pathogenesis of invasive transitional cell carcinoma. Protein homology analysis of CAR revealed that it is a member of the Ig-type superfamily of cell-cell adhesion molecules.25 These cellcell adhesion molecules share common features, such as the evolutionarily conserved Ig C2 type domains, which have a relatively rigid structure and allow these molecules to project out of the cell membrane. The cytoplasmic domains of these various adhesion molecules may interact with structural elements of the cell (eg, the catenins), as well as regulatory mechanisms. The primary binding mode of the Ig family of cell-cell adhesion molecules is homophilic self-association; the integrins and selectins recognize other types of molecules in a heterophilic-binding mode. Interestingly, bladder cancer cells have long been known to desquamate more readily in high-grade, invasive disease. It may well be that part of this tendency to lose cohesiveness has to do with the loss of CAR expression. This hypothesis was recently supported by Okegawa et al.6 who showed that CAR facilitates intercellular adhesion and can inhibit cell growth. 535
CONCLUSIONS The results of this study demonstrate for the first time that CAR and integrin ␣v are inversely expressed in clinical specimens of bladder cancer. CAR down-regulation would render invasive bladder cancer a poor target for adenoviral gene therapy and could only be overcome by either changing the tropism of adenovirus or by up-regulating the receptor. However, we also hypothesize that an induction of the CAR expression would affect not only the susceptibility of bladder cancer to adenoviral gene therapy but also perhaps the natural history of bladder cancer in general. This finding could have profound implications in the development of intravesical therapies for carcinoma in situ and high-grade T1 lesions, both of which often require radical surgery when conventional intravesical therapy fails. Recent evidence has demonstrated that CAR expression can often be re-established through the use of certain histone deacetylase inhibitors.26 An evaluation of potential CAR re-expression in bladder cancer cells by such HDAI is warranted on the basis of these preliminary results. Influencing the CAR expression pattern could also change the pathogenesis of such lesions by preventing the development of the invasive phenotype or even reverting tumor invasion. ACKNOWLEDGMENT. To Onyx Pharmaceuticals for generously providing us with the proprietary polyclonal antibody CAR 72 and Jason Le for expert technical assistance; to Wasim Chowdhury for his excellent help with designing the figures; and to William Nelson, Seth Lerner, Robert Getzenberg, and J. T. Hsieh for many helpful suggestions and discussions regarding this work. REFERENCES 1. Greenlee RT, Hill-Harmond MB, Murray T, et al: Cancer Statistics 2001. CA Cancer J Clin 51: 15–36, 2001. 2. Morris BD Jr, Drazan KE, Csete ME, et al: Adenoviralmediated gene transfer to bladder in vivo. J Urol 152: 506 – 509, 1994. 3. Bergelson JM, Cunningham JA, Droguett G, et al: Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275: 1320 –1323, 1997. 4. Nemerow GR, and Stewart PL: Role of alpha(v) integrins in adenovirus cell entry and gene delivery. Microbiol Mol Biol Rev 63: 725–734, 1999. 5. Honda T, Saitoh H, Masuko M, et al: The coxsackievirus-adenovirus receptor protein as a cell adhesion molecule in the developing mouse brain. Brain Res Mol Brain Res 77: 19 – 28, 2000. 6. Okegawa T, Pong RC, Li Y, et al: The mechanism of the growth-inhibitory effect of coxsackie and adenovirus receptor (CAR) on human bladder cancer: a functional analysis of CAR protein structure. Cancer Res 61: 6592– 6600, 2001. 7. Johnstone CN, Tebbutt NC, Abud HE, et al: Characterization of mouse A33 antigen, a definitive marker for basolat-
536
eral surfaces of intestinal epithelial cells. Am J Physiol Gastrointest Liver Physiol 279: G500 –G510, 2000. 8. Palmeri D, van Zante A, Huang CC, et al: Vascular endothelial junction-associated molecule, a novel member of the immunoglobulin superfamily, is localized to intercellular boundaries of endothelial cells. J Biol Chem 275: 19139 –19145, 2000. 9. Ruoslahti E, and Pierschbacher MD: New perspectives in cell adhesion: RGD and integrins. Science 238: 491– 497, 1987. 10. Hynes RO: Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69: 11–25, 1992. 11. Tamkun JW, DeSimone DW, Fonda D, et al: Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46: 271–282, 1986. 12. Kornberg LJ, Earp HS, Turner CE, et al: Signal transduction by integrins: increased protein tyrosine phosphorylation caused by clustering of beta 1 integrins. Proc Natl Acad Sci USA 88: 8392– 8396, 1991. 13. Brooks PC, Clark RA, and Cheresh DA: Requirement of vascular integrin alpha v beta 3 for angiogenesis. Science 264: 569 –571, 1994. 14. Brooks PC, Stromblad S, Sanders LC, et al: Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alpha v beta 3. Cell 85: 683– 693, 1996. 15. Jones J, Watt FM, and Speight PM: Changes in the expression of alpha v integrins in oral squamous cell carcinomas. J Oral Pathol Med 26: 63– 68, 1997. 16. Sheppard D: Epithelial integrins. Bioassays 18: 655– 660, 1996. 17. Miller SE, and Veale RB: Environmental modulation of alpha(v), alpha(2) and beta(1) integrin subunit expression in human oesophageal squamous cell carcinomas. Cell Biol Int 25: 61– 69, 2001. 18. Bello L, Francolini M, Marthyn P, et al: Alpha(v)beta3 and alpha(v)beta5 integrin expression in glioma periphery. Neurosurgery 49: 380 –389, 2001. 19. Wechsel HW, Petri E, Feil G, et al: Renal cell carcinoma: immunohistological investigation of expression of the integrin alpha v beta 3. Anticancer Res 19: 1529 –1532, 1999. 20. Takayama K, Ueno H, Pei XH, et al: The levels of integrin alpha v beta 5 may predict the susceptibility to adenovirus-mediated gene transfer in human lung cancer cells. Gene Ther 5: 361–368, 1998. 21. Li Y, Pong RC, Bergelson JM, et al: Loss of adenoviral receptor expression in human bladder cancer cells: a potential impact on the efficacy of gene therapy. Cancer Res 59: 325– 330, 1999. 22. Kononen J, Bubendorf L, Kallioniemi A, et al: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4: 844 – 847, 1998. 23. Rauen KA, Sudilovsky D, Le J, et al: Expression of the coxsackie-adenovirus receptor (CAR) in normal prostate and in primary and metastatic prostate carcinoma: potential relevance to gene therapy. Cancer Res 62: 3812–3818, 2002. 24. Sutton MA, Berkman SA, Chen SH, et al: Adenovirusmediated suicide gene therapy for experimental bladder cancer. Urology 49: 173–180, 1997. 25. Ruoslahti E, and Obrink B: Common principles in cell adhesion. Exp Cell Res 227: 1–11, 1996. 26. Kitazono M, Goldsmith ME, Aikou T, et al: Enhanced adenovirus transgene expression in malignant cells treated with the histone deacetylase inhibitor FR901228. Cancer Res 61: 6328 – 6330, 2001.
UROLOGY 60 (3), 2002
INTEGRIN ␣v AND COXSACKIE ADENOVIRUS RECEPTOR EXPRESSION IN CLINICAL BLADDER CANCER MARKUS D. SACHS, KATHERINE A. RAUEN, MEERA RAMAMURTHY, JENNIFER L. DODSON, ANGELO M. DE MARZO, MATHEW J. PUTZI, MARK P. SCHOENBERG, AND RONALD RODRIGUEZ
ABSTRACT Objectives. To evaluate the expression of the coxsackie and adenovirus receptor (CAR) and ␣v integrins in clinical specimens of bladder cancer to determine the susceptibility to adenoviral gene therapy. Efficient adenovirus-based gene therapy requires binding of the virus to CAR and involves the ␣v integrins. Studies on bladder cancer cell lines have shown that low adenoviral transduction rates were associated with low-level expression of CAR. Integrin ␣v expression increases in various tumors suggest its importance in differentiation, proliferation, and migration. CAR is structurally a member of the Ig-type superfamily of cell-cell adhesion molecules, suggesting that its expression may also be related to the state of tumor differentiation. Methods. We performed immunohistochemistry for CAR and integrin ␣v expression in bladder cancer specimens in 50 paraffin-embedded tumor-normal pairs and confirmed the results by quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of 11 separate bladder tumors and 4 separate normal bladder controls. Results. Immunochemistry demonstrated a stage and grade-dependent decrease in CAR expression (90.0%, 83.3%, and 31.3% of normal urothelium and superficial and invasive transitional cell carcinoma [TCC] and 83.3% and 39.5% of low and high-grade TCC, respectively). Furthermore, we found a stage and grade-dependent increase in ␣v integrin expression (13.3%, 46.0%, and 56.3% of normal urothelium, superficial TCC, and invasive TCC and 25% and 52.6% of low and high-grade TCC, respectively). Quantitative RT-PCR analysis confirmed a downregulation at the CAR gene expression level. Conclusions. This down-regulation may have a major impact on developing adenoviral-based gene therapy modalities. In addition, we propose that loss of CAR expression decreases rigid cell adhesion, possibly increasing the migratory potential. Loss of CAR expression correlates with the invasive phenotype in our analysis of bladder cancer. Simultaneously, the finding of increased ␣v expression in invasive cancer suggests a pathogenesis that involves heterophilic adhesion and migration of these cells on various extracellular ligands. UROLOGY 60: 531–536, 2002. © 2002, Elsevier Science Inc.
C
arcinoma of the bladder is the second most common cancer of the urogenital tract and makes up approximately 6% of all malignancies. In This work was supported in part by a research grant to M. D. Sachs from the Max Kade Foundation, New York, New York. From the Brady Urological Institute, Johns Hopkins Hospital, Baltimore, Maryland; Department of Urology, Charite Medical School, Humboldt University, Berlin, Germany; Department of Pediatrics, University of California, San Francisco, School of Medicine; Cancer Research Institute, Comprehensive Cancer Center, San Francisco, California; and Johns Hopkins Oncology Center, Department of Surgical Pathology, Johns Hopkins Hospital, Baltimore, Maryland Reprint requests: Ronald Rodriguez, M.D., Ph.D., Brady Urological Institute, Marburg 205, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287 Submitted: December 7, 2001, accepted (with revisions): April 1, 2002 © 2002, ELSEVIER SCIENCE INC. ALL RIGHTS RESERVED
2001, an estimated 54,300 people will have been diagnosed and an estimated 12,400 patients will have died of bladder cancer in the United States.1 Bladder cancer treatment failure is a morbid condition with limited options. Gene therapy directed against these failures represents a novel and evolving alternative.2 Adenovirus binding and uptake are separate but cooperative events. Binding requires interaction between the knob domain of the adenoviral fiber protein present at each vertex of the icosahedral structure of the virus and the coxsackie and adenovirus receptor (CAR),3 and internalization involves a secondary interaction between the penton base and certain patterns of integrins, in particular, ␣v3 and ␣v5.4 CAR is a 46-kD, class I, transmembrane protein with two Ig-like C2 domains. Struc0090-4295/02/$22.00 PII S0090-4295(02)01748-X 531
turally it is a member of the Ig superfamily of adhesion molecules and under certain conditions has been shown to suppress cell growth in vitro.5,6 CAR has a structure highly homologous to other known adhesion molecules, such as A33 antigen7 and vascular endothelial junction-associated protein.8 Integrins are cell-surface adhesion molecules that regulate normal cellular interactions, such as cell migration and assembly of, and interaction with, extracellular matrix ligands (eg, vitronectin) by binding Arg-Gly-Asp (RGD) amino acid sequences,9 cell growth and differentiation, and cell survival. These heterodimeric integral membrane proteins are composed of an ␣-chain and a -chain and facilitate signal transduction in a bidirectional manner, thereby having a direct impact on cell behavior.10 –12 Aberrant integrin expression is believed to play a role in tumor invasion and metastasis. Integrin ␣v plays a variety of roles in carcinogenesis, including differentiation, proliferation, prevention of apoptosis, matrix degradation, and adhesion.13–16 De novo ␣v expression in esophageal squamous cell carcinoma has been suggested to decrease rigid homophilic cell adhesion, possibly increasing the migratory potential while simultaneously permitting the heterophilic adhesion and migration of malignant cells on various ligands.17 Integrins ␣v3 and ␣v5 play a role in glioma-associated angiogenesis, neoangiogenesis, and cell migration in high-grade gliomas.18 Wechsel et al.19 described a grade-dependent increase in ␣v3 expression in renal cell carcinoma as a step in metastasis formation by increasing adhesion to the basement membrane. Although adenoviral binding requires an interaction with CAR, internalization is at least partly mediated by an interaction between the viral penton base and the ␣v3 or ␣v5 integrins. Takayama et al.20 found a correlation between the level of ␣v5 integrin and the expression level of a transferred gene and suggested that quantification of ␣v5 expression may be a good way of predicting the susceptibility of cells to adenoviral infection. In bladder cancer cell lines, the binding of adenovirus by way of CAR seems to be impaired. Li et al.21 reported that several human bladder cancer cell lines lacked expression of CAR, although the genomic sequences for the gene were intact. This downregulation correlated with resistance of urothelial cells to adenoviral infection.21 More recently, Okegawa et al.6 showed an apparent stagedependent reduction of CAR mRNA in clinical bladder cancer specimens using a semiquantitative polymerase chain reaction (PCR) method. A reduction in the actual CAR protein expression in bladder cancer, if proved, would have a major impact on the development of new adenoviral-based gene therapies. We present a detailed analysis of CAR and inte532
grin expression performed on bladder cancer specimens. In addition to predicting the clinical susceptibility to adenoviral gene transfer, these data may also provide valuable insight into the development of the invasive phenotype in bladder cancers. MATERIAL AND METHODS CELL LINE HEK293 cells (human embryo kidney, a packaging cell line for adenovirus, known to express very high levels of CAR) were purchased from American Tissue Culture Collection (Manassas, Va). Cells were grown in Dulbecco’s modification of Eagle’s medium (Mediatech), supplemented with 10% fetal bovine serum and 0.1% gentamicin.
TISSUE SAMPLES Normal and cancerous paraffin-embedded formalin-fixed bladder specimens from 50 patients (40 men and 10 women) were obtained from the pathology archives of Johns Hopkins Hospital for immunohistochemical analysis. Tumor samples were available from all 50 patients; normal urothelium was only available from 30 patients. Permission to evaluate these tissues was obtained from our institutional review board. Tissue sections were arranged in a tissue microarray, as previously described, using 0.6-mm cores.22 The mean patient age was 63.7 ⫾ 9.5 years (range 41 to 77). Tumor stage was assigned according to the American Joint Committee on Cancer staging 1997 TNM system. The pathologic stage was as follows: pTa, n ⫽ 7; pTis, n ⫽ 5; pT1, n ⫽ 6; pT2, n ⫽ 6; pT3, n ⫽ 16; and pT4, n ⫽ 10. Snap-frozen additional specimens of normal urothelium (n ⫽ 4) and bladder carcinoma Stage Tis (n ⫽ 1), Ta (n ⫽ 6), T1 (n ⫽ 2), T2 (n ⫽ 1), and T3 (n ⫽ 1) were obtained for quantitative reverse transcriptase (RT)-PCR analysis.
IMMUNOHISTOCHEMISTRY FOR CAR AND ␣V Integrin Expression Tissue sections were deparaffinized in xylene and dehydrated in ethanol. For CAR staining, standard indirect immunoperoxidase methods were used for immunostaining with rabbit polyclonal CAR 72 (Onyx Pharmaceuticals, Richmond, Calif), as previously described.23 In brief, the slides were baked for 30 minutes at 60°C and standard antigen retrieval methods, including trypsinization and microwave treatment in 10 mM citrate buffer, were performed. The tissue was blocked in 10% goat serum/phosphate-buffered saline and incubated in primary polyclonal antibody CAR 72 diluted 1:7000 for 8 to 16 hours at 4°C. The tissue sections were incubated in secondary biotinylated goat anti-rabbit immunoglobulin diluted 1:200 (Vector Laboratories, Burlingame, Calif) and then treated with streptavidin-biotinylated horseradish peroxidase complex (Vectastain Elite ABC Kit, Vector Laboratories). The sections were subsequently developed using diaminobenzidine tetrahydrochloride (Sigma) in hydrogen peroxide/phosphate-buffered saline and counterstained with hematoxylin. Staining for integrin ␣v was done using a monoclonal antibody (P2W7, Santa Cruz Biotechnology, Santa Cruz, Calif) at a 1:50 dilution with an overnight incubation and the LSAB2 System-HRP (Dako, Carpinteria, Calif). Both stainings were characterized in five staining levels (0, no staining; 1, 1% to 25% of cells showing membrane bound expression; 2, 25% to 50%; 3, 50% to 75%; and 4, 75% to 100%) by a single uropathologist (A.M.D.). For the purposes of presentation, the staining result was considered positive if more than 50% of cells expressed the antigen (staining levels 3 and 4) and negative if 50% or less of the cells showed a staining signal (levels 0 to 2). The staining is shown in Figure 1. UROLOGY 60 (3), 2002
denaturing for 15 seconds at 94°C, annealing for 30 seconds at 59.6°C and extending for 45 seconds at 70°C for 38 cycles, followed by a 7-minute final extension at 72°C.
STATISTICAL ANALYSIS The immunohistochemistry results were analyzed in relation to tumor stage and grade using the Kruskal-Wallis test for the average staining level and the chi-square test for the staining results (positive for staining levels 0 to 2 and negative for staining levels 3 to 4). Staining of tumor versus normal tissue was analyzed using the chi-square test. A result was considered statistically significant when the P value was less than 0.05. Statistical analysis was done using StatView for Windows, version 5.0.
RESULTS
FIGURE 1. CAR and integrin ␣v immunohistochemistry. (A) CAR-positive normal urothelium. (B) CAR-positive noninvasive transitional cell carcinoma. (C) CAR-negative invasive transitional cell carcinoma. (D) ␣v-negative normal urothelium. (E) ␣v-negative noninvasive transitional cell carcinoma. (F) ␣v-positive invasive transitional cell carcinoma (original magnification ⫻ 400).
REAL-TIME QUANTITATIVE RT-PCR FOR CAR RNA isolates were prepared from HEK293 cells and from fresh frozen tissue (normal bladder controls, n ⫽ 4, and bladder tumors, n ⫽ 11), using MicroPoly(A) Pure Kit (Ambion, Austin, Tex). PCRs were set up in a reaction volume of 50 L containing 23 L of reaction mix (Life Technologies, Rockville, Md), 8 L of 5 mM magnesium sulfate (Life Technologies), 10 L RNA, 2 L of each 5 M amplification primer, 1 L of 1:25,000 Sybr Green (Roche, Indianapolis, Ind), and 1 L of rt-Taq (Life Technologies). The sense and antisense primer sequence was 5⬘-TTGCTTGCTCTAGCGCTCATTGGTC-3⬘ and 5⬘-TCATCACAGGAATCGCACCCATTCG-3⬘, respectively (product size 325 bp). The amounts of target DNA (relative CAR copy number per haploid genome) were quantified using the following equation: Relative CAR copy number per haploid genome ⫽
CAR copy number . -actin copy number
The CAR copy number was calculated by generating a standard curve based on the amplification of known amounts of a plasmid carrying the CAR sequence. RNA values were then normalized by amplifying the constitutively expressed housekeeping gene -actin. The -actin copy number was calculated using serial dilutions of a known amount of -actin mRNA. Reactions were carried out in the iCycler thermal cycler (Bio-Rad, Hercules, Calif). The thermal cycling protocol was 30 minutes at 50°C and 2 minutes at 94°C, with subsequent UROLOGY 60 (3), 2002
INTEGRIN ␣V AND CAR PROTEIN EXPRESSION A summary of the staining results and staining levels is given in Figures 1 and 2. Staining levels 1 through 5 were assigned according to the percentage of cells that showed membranous staining (see Material and Methods section). To assess the staining results, we defined tumors as positive if more than 50% of cells showed a membranous staining signal (staining levels 3 and 4). Because this is a subjective definition, the actual staining level is also shown in a graph (Fig. 2C,D). Using these criteria, 50.0% (25 of 50) of the bladder tumors examined were considered negative for CAR expression compared with 10% (3 of 30) of the normal urothelium (obtained from the same bladder specimen). This downregulation was dependent on the tumor stage and grade, with 83.3% (15 of 18) versus 31.3% (10 of 32) expression in superficial (Ta, T1, Tis) and invasive (T2-T4) tumors, respectively, and 83.3% (10 of 12) versus 39.5% (15 of 38) in well and poorly differentiated tumors, respectively (Fig. 2A,B). The results were statistically significant when assessing the average staining level (0 to 4) and tumor stage (P ⫽ 0.0014, Kruskal-Wallis test), staining result (positive or negative) and tumor stage (P ⫽ 0.013, chi-square test), average staining level and tumor grade (P ⫽ 0.028, Kruskal-Wallis test), and staining result and tumor grade (P ⫽ 0.049, chi-square test). CAR expression in normal urothelium was equal in umbrella cells and more basal layers. CAR staining was positive in 90.0% of normal urothelium compared with in 50% of tumors, which was statistically significant (P ⫽ 0.0005, chi-square test). No significant differences were found between normal urothelial components from samples of the various stages and grades of bladder cancer specimens. For ␣v integrin, we found an overall positive staining of 46.0% (23 of 50) in bladder tumors and 13.3% (4 of 30) of normal urothelium. This difference was statistically significant (P ⫽ 0.003, chisquare test). The increase in ␣v expression was stage and grade dependent, with 56.3% (18 of 32) 533
FIGURE 2. CAR and integrin ␣v expression in bladder tumors. (A) Percentage of tumors with CAR staining level 3 or 4 (more than 50% of cells positive) and tumor stage and grade. (B) Average CAR staining level and tumor stage and grade. (C) Percentage of tumors with integrin ␣v staining level 3 or 4 and tumor stage and grade. (D) Average integrin ␣v staining level and tumor stage and grade.
of invasive tumor samples versus 27.8% (5 of 18) of superficial tumor samples and 52.6% (20 of 38) of high-grade tumor samples versus 25% (3 of 12) of low-grade tumor samples (Figure 2C,D) showing an increase. However, these values did not reach statistical significance (P ⫽ 0.234, chi-square test for comparison with stage, and P ⫽ 0.246, chi-square test for comparison with grade). The integrin ␣v staining in normal urothelium was more prominent in umbrella cells than in more basal layers. Stromal cells were universally negative for both CAR and integrin ␣v. REAL-TIME PCR Because the CAR antibody is an unvalidated, noncommercial reagent, we confirmed the CAR immunohistochemistry findings with real-time quantitative RT-PCR in a separate subset of samples obtained prospectively. As found in the immunohistochemistry results, a marked down-regulation of CAR occurred in bladder tumors (n ⫽ 11) compared with normal urothelium (n ⫽ 4), supporting the finding that CAR expression is downregulated in invasive bladder cancer. The HEK293 cells served as a positive control (Fig. 3). 534
COMMENT Because CAR and ␣v integrins are important for adenoviral binding and internalization, we sought to evaluate their expression patterns in the same samples of clinical bladder cancer. Interestingly, we found that the more invasive cancers tended to have an increased ␣v integrin pattern and a decreased CAR pattern of expression. The CAR expression data were more dramatic and easily reached statistical significance for this relatively small sample size. Although the overall difference of ␣v integrin expression between tumor and normal tissue was statistically significant, no significant difference was noted when segregated by stage and grade. Nonetheless, there clearly appeared to be a trend toward increased ␣v expression with stage and grade, which requires larger sample numbers to better assess. Recent analysis of bladder cancer cell lines has demonstrated marked down-regulation of CAR expression, as well as a significant resistance to transgene expression by adenoviral vectors.21 In addition, preclinical models of bladder cancer gene therapy have demonstrated an unusually high requirement for adenoviral vectors.24 This raised the UROLOGY 60 (3), 2002
FIGURE 3. Evaluation of CAR expression by quantitative RT-PCR. Average copy number with standard error.
question of whether adenoviral-based gene transduction into human bladder cancer cells would be possible. Previously, CAR expression in clinical bladder cancer samples had only been analyzed semiquantitatively by mRNA expression, suggesting an apparent down-regulation in the invasive tumors.6 We present a larger, more detailed analysis based on the protein expression pattern. Moreover, without an immunohistologic correlation, it is also unknown whether this down-regulation applies to all urothelial cells or to subpopulations. A reduction in active receptors would clearly lead to decreased viral gene delivery. Because CAR seems to play a role in cell-cell adhesion and migration, such a reduction might cause cells to lose cohesiveness and therefore may be a major contributing factor in the process of developing the invasive phenotype. We demonstrated by immunohistochemistry that clinical specimens of invasive bladder cancers reveal a marked stage and grade-dependent downregulation of CAR. Carcinoma in situ was evaluated as a superficial cancer, although one should bear in mind that this is a poorly differentiated tumor with high invasive potential. However, if carcinoma in situ is taken out of the data analysis, neither the overall rate (46.7%) nor the CAR positive rate of superficial tumors (84.6%) changed dramatically. Although the number of available samples was still low, these data suggest that CAR expression varies with the invasive potential of the cancer cells in clinical specimens of bladder cancer. We independently confirmed the marked downregulation of CAR in bladder tumors compared with normal urothelium by real-time quantitative RT-PCR. UROLOGY 60 (3), 2002
Our findings have two major clinical impacts. First, by characterizing the adenoviral receptor status as a function of clinical stage and grade, rational targeting of the appropriate patient populations suitable for gene therapy trials can be predicted. Moreover, for those patients who lack the appropriate receptor status, either modification to the adenoviral vectors will be required to allow high efficiency gene transfer or CAR expression will need to be reconstituted. Gene transfer with nonadenovirus vectors might be an alternative approach in patients with low CAR-expressing tumors. Second, these results provide important information about the pathogenesis of invasive transitional cell carcinoma. Protein homology analysis of CAR revealed that it is a member of the Ig-type superfamily of cell-cell adhesion molecules.25 These cellcell adhesion molecules share common features, such as the evolutionarily conserved Ig C2 type domains, which have a relatively rigid structure and allow these molecules to project out of the cell membrane. The cytoplasmic domains of these various adhesion molecules may interact with structural elements of the cell (eg, the catenins), as well as regulatory mechanisms. The primary binding mode of the Ig family of cell-cell adhesion molecules is homophilic self-association; the integrins and selectins recognize other types of molecules in a heterophilic-binding mode. Interestingly, bladder cancer cells have long been known to desquamate more readily in high-grade, invasive disease. It may well be that part of this tendency to lose cohesiveness has to do with the loss of CAR expression. This hypothesis was recently supported by Okegawa et al.6 who showed that CAR facilitates intercellular adhesion and can inhibit cell growth. 535
CONCLUSIONS The results of this study demonstrate for the first time that CAR and integrin ␣v are inversely expressed in clinical specimens of bladder cancer. CAR down-regulation would render invasive bladder cancer a poor target for adenoviral gene therapy and could only be overcome by either changing the tropism of adenovirus or by up-regulating the receptor. However, we also hypothesize that an induction of the CAR expression would affect not only the susceptibility of bladder cancer to adenoviral gene therapy but also perhaps the natural history of bladder cancer in general. This finding could have profound implications in the development of intravesical therapies for carcinoma in situ and high-grade T1 lesions, both of which often require radical surgery when conventional intravesical therapy fails. Recent evidence has demonstrated that CAR expression can often be re-established through the use of certain histone deacetylase inhibitors.26 An evaluation of potential CAR re-expression in bladder cancer cells by such HDAI is warranted on the basis of these preliminary results. Influencing the CAR expression pattern could also change the pathogenesis of such lesions by preventing the development of the invasive phenotype or even reverting tumor invasion. ACKNOWLEDGMENT. To Onyx Pharmaceuticals for generously providing us with the proprietary polyclonal antibody CAR 72 and Jason Le for expert technical assistance; to Wasim Chowdhury for his excellent help with designing the figures; and to William Nelson, Seth Lerner, Robert Getzenberg, and J. T. Hsieh for many helpful suggestions and discussions regarding this work. REFERENCES 1. Greenlee RT, Hill-Harmond MB, Murray T, et al: Cancer Statistics 2001. CA Cancer J Clin 51: 15–36, 2001. 2. Morris BD Jr, Drazan KE, Csete ME, et al: Adenoviralmediated gene transfer to bladder in vivo. J Urol 152: 506 – 509, 1994. 3. Bergelson JM, Cunningham JA, Droguett G, et al: Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 275: 1320 –1323, 1997. 4. Nemerow GR, and Stewart PL: Role of alpha(v) integrins in adenovirus cell entry and gene delivery. Microbiol Mol Biol Rev 63: 725–734, 1999. 5. Honda T, Saitoh H, Masuko M, et al: The coxsackievirus-adenovirus receptor protein as a cell adhesion molecule in the developing mouse brain. Brain Res Mol Brain Res 77: 19 – 28, 2000. 6. Okegawa T, Pong RC, Li Y, et al: The mechanism of the growth-inhibitory effect of coxsackie and adenovirus receptor (CAR) on human bladder cancer: a functional analysis of CAR protein structure. Cancer Res 61: 6592– 6600, 2001. 7. Johnstone CN, Tebbutt NC, Abud HE, et al: Characterization of mouse A33 antigen, a definitive marker for basolat-
536
eral surfaces of intestinal epithelial cells. Am J Physiol Gastrointest Liver Physiol 279: G500 –G510, 2000. 8. Palmeri D, van Zante A, Huang CC, et al: Vascular endothelial junction-associated molecule, a novel member of the immunoglobulin superfamily, is localized to intercellular boundaries of endothelial cells. J Biol Chem 275: 19139 –19145, 2000. 9. Ruoslahti E, and Pierschbacher MD: New perspectives in cell adhesion: RGD and integrins. Science 238: 491– 497, 1987. 10. Hynes RO: Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69: 11–25, 1992. 11. Tamkun JW, DeSimone DW, Fonda D, et al: Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin. Cell 46: 271–282, 1986. 12. Kornberg LJ, Earp HS, Turner CE, et al: Signal transduction by integrins: increased protein tyrosine phosphorylation caused by clustering of beta 1 integrins. Proc Natl Acad Sci USA 88: 8392– 8396, 1991. 13. Brooks PC, Clark RA, and Cheresh DA: Requirement of vascular integrin alpha v beta 3 for angiogenesis. Science 264: 569 –571, 1994. 14. Brooks PC, Stromblad S, Sanders LC, et al: Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alpha v beta 3. Cell 85: 683– 693, 1996. 15. Jones J, Watt FM, and Speight PM: Changes in the expression of alpha v integrins in oral squamous cell carcinomas. J Oral Pathol Med 26: 63– 68, 1997. 16. Sheppard D: Epithelial integrins. Bioassays 18: 655– 660, 1996. 17. Miller SE, and Veale RB: Environmental modulation of alpha(v), alpha(2) and beta(1) integrin subunit expression in human oesophageal squamous cell carcinomas. Cell Biol Int 25: 61– 69, 2001. 18. Bello L, Francolini M, Marthyn P, et al: Alpha(v)beta3 and alpha(v)beta5 integrin expression in glioma periphery. Neurosurgery 49: 380 –389, 2001. 19. Wechsel HW, Petri E, Feil G, et al: Renal cell carcinoma: immunohistological investigation of expression of the integrin alpha v beta 3. Anticancer Res 19: 1529 –1532, 1999. 20. Takayama K, Ueno H, Pei XH, et al: The levels of integrin alpha v beta 5 may predict the susceptibility to adenovirus-mediated gene transfer in human lung cancer cells. Gene Ther 5: 361–368, 1998. 21. Li Y, Pong RC, Bergelson JM, et al: Loss of adenoviral receptor expression in human bladder cancer cells: a potential impact on the efficacy of gene therapy. Cancer Res 59: 325– 330, 1999. 22. Kononen J, Bubendorf L, Kallioniemi A, et al: Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4: 844 – 847, 1998. 23. Rauen KA, Sudilovsky D, Le J, et al: Expression of the coxsackie-adenovirus receptor (CAR) in normal prostate and in primary and metastatic prostate carcinoma: potential relevance to gene therapy. Cancer Res 62: 3812–3818, 2002. 24. Sutton MA, Berkman SA, Chen SH, et al: Adenovirusmediated suicide gene therapy for experimental bladder cancer. Urology 49: 173–180, 1997. 25. Ruoslahti E, and Obrink B: Common principles in cell adhesion. Exp Cell Res 227: 1–11, 1996. 26. Kitazono M, Goldsmith ME, Aikou T, et al: Enhanced adenovirus transgene expression in malignant cells treated with the histone deacetylase inhibitor FR901228. Cancer Res 61: 6328 – 6330, 2001.
UROLOGY 60 (3), 2002