Combination Analysis of Activator Protein-1FamilyMembers, Sp1and an Activator Protein-2A-Related Factor Binding to Different Regions of the Urokinase Receptor Gene in Resected Colorectal Cancers

Purpose: Studies on the transactivation of genes via promoter elements have mostly been done on cell lines rather than resected tissues. This, however, is essential to address an in vivo or clinical relevance.We have previously shown tumor-specific binding of Sp1and an activator protein (AP)-2^ related factor to promoter region 152/ 135 of the metastasis-related u-PAR gene in 60% of in vivo ^ resected cancer tissues. Cell lines have implicated an additional role, and potential synergism, of an AP-1 region ( 190/ 171) in u-PAR regulation. This study was done to (a) analyze AP-1 binding to this region in resected tumor and normal tissues, and define subgroups in which it is tumor-specific, and (b) to analyze transcription factor ^binding patterns to both promoter motifs in resected tissues, supporting synergism, and draw first prognostic conclusions. Experimental Design: In 103 patients with colorectal cancer, electrophoretic mobility shift assay/supershift analysis for u-PAR promoter region 190/ 171 was done in tumors and normal tissues. In 71 patients, region 152/ 135 was also analyzed. U-PAR protein was measured by ELISA. Results:Tumor-specific AP-1binding to region 190/ 171of the u-PAR promoter was found in 40% of patients. Subgroup analysis showed tumor-specific binding for c-Fos in 58%, for c-Jun in 50%, for JunD in 39%, and for Fra-1 in 4% of cases. AP-1binding correlated significantly with u-PAR protein amounts in both normal and tumor tissues (P < 0.001), in contrast to a tumor-specific correlation with u-PAR of the AP-2/Sp1 region. In analyses for both promoter regions, 62% of cancers showed simultaneous binding for AP-1, AP-2, and Sp1, 11% for AP-1 and AP-2, 16% for AP-2 and Sp1, 4% for AP-2 only, 3% for AP-1 only, and 0% for Sp1 only. The binding of AP-1, AP-2, and Sp1 correlated significantly with each other (P < 0.001), the combination of AP-1 and AP-2 showing the highest correlation with u-PAR (P = 0.008). Preliminary survival analysis indicated a trend for poorer prognosis for binding of all three transcription factors. Conclusion:This is the first study differentiating transcription factor ^ binding to two important u-PAR promoter regions in a large series of resected tumors and normal tissues. The AP-1 site seems to be a less tumor-specific regulator than the Sp1/AP-2 motif. Nevertheless, data corroborate the hypothesis of synergism between both elements in resected tumors. Analyses of molecular pathways have largely been attempted in artificial cell line models which have acquired several molecular and phenotypic changes during culture. Therefore, their ability to reflect the situation in naturally occurring tumors, or tissues in general, has to be questioned. Specifically, research on transcriptional regulation has rarely been done in authentic resected Human Cancer Biology Authors’Affiliations: DepartmentofExperimentalSurgeryandMolecularOncology, UniversitaetsklinikumMannheim and Department of Surgery Mannheim, University ofHeidelberg,Heidelberg, Departmentof Pediatrics, Dr. v. HaunerschesKinderspital, Klinikum Grosshadern, Ludwig Maximilians University, Munich, Germany, and Department ofObstetrics andGynecology, University of Chicago, Chicago, Illinois Received 4/20/05; revised 9/30/05; accepted10/4/05. Grant support: Wilhelm Sander Stiftung, Munich, the Auguste SchaedelDantscher-Stiftung, Garmisch, and the Alfried Krupp Prize for Young University Teachers of the Krupp Foundation, Essen, Germany (H. Allgayer). The costs of publication of this article were defrayed in part by the payment of page charges.This article must therefore be hereby marked advertisement in accordance with18U.S.C. Section1734 solely to indicate this fact. Note: D.M. Schewe and T. Biller share first authorship. This publication contains parts of the disser tation of D.M. Schewe and T. Biller performed in partial fulfillment of the requirements for the ‘‘Dr. med’’ at the Faculty of Medicine, Ludwig Maximilians University, Munich, and the Mannheim Faculty of Medicine, University of Heidelberg, Germany, respectively. Requests for reprints: Heike Allgayer, Department of Experimental Surgery/ Molecular Oncology, Universitätsklinikum Mannheim, University of Heidelberg, 68167 Mannheim, Germany. Phone: 49-621-383-2226; Fax: 49-621-3833809/1938; E-mail: [email protected]. F2005 American Association for Cancer Research. doi:10.1158/1078-0432.CCR-05-0786 www.aacrjournals.org Clin Cancer Res 2005;11(24) December15, 2005 8538 Research. on April 12, 2017. © 2005 American Association for Cancer clincancerres.aacrjournals.org Downloaded from tumor material from a relevant number of patients. However, such analyses will be increasingly important when it comes to applying new targeting strategies to the clinical setting, for which an appropriate patient selection will be essential. The urokinase receptor (u-PAR) is a Mr 55,000 to 60,000 heavily glycosylated, disulfide-linked cell surface receptor leading to plasmin-mediated degradation of extracellular matrix components such as fibrin and collagen IV, this being achieved by its ligand urokinase-type plasminogen activator (u-PA; refs. 1–8). As a central molecule to promote invasion and metastasis, it has been shown to be overexpressed and associated with a poor prognosis in diverse human tumors such as breast and especially gastrointestinal cancers (9–17). Currently, evidence is sufficient to support the concept that a high u-PAR expression in cancer is largely brought about by transcriptional activation of the gene, although further mechanisms such as mRNA-destabilizing motifs, posttranslational modifications, and endocytotic cleavage are additional means of controlling u-PAR protein amounts (2, 5, 18–21). Our previous work in cultured colon cancer (22, 23) had shown that a motif spanning region 152/ 135 of the u-PAR promoter, bound with Sp1, Sp3, and an activator protein (AP)-2a–related protein, was one of the essential mediators of a high constitutive as well as phorbol 12-myristate 13-acetate– and c-src –inducible u-PAR gene expression. These results in cell lines could be confirmed in vivo in our first study (24), in which we had analyzed the relevance and distribution of transcription factor binding to this promoter element in the largest published series thus far of 145 gastrointestinal carcinomas and corresponding normal tissues. AP-2 and Sp1 binding correlated significantly with high u-PAR amounts in tumor tissues in contrast to normal mucosa, implicating that regulation of u-PAR gene expression via this motif might become biologically relevant in the transition to the malignant state. Most importantly, a tumor-specific transcription factor binding of Sp1 and the AP-2a-related protein was observed in up to 60% of the patients, suggesting u-PAR promoter region 152/ 135, and molecular pathways mediated by it, as promising tumorselective targets for countering u-PAR-mediated invasion and metastasis in this large subgroup of gastrointestinal cancer patients. An additional promoter region that is essential for u-PAR gene expression in cultured colon cancer had been defined by Lengyel et al. (25), spanning base pairs 190/ 171 and an AP-1consensusmotif. In vitro studies in colon cancer have shown that constitutive and phorbol 12-myristate 13-acetate–inducible gene expression required this promoter region bound with the AP-1 family members Jun-D, c-Jun , c-Fos , and Fra-1 (25). This motif also mediated an induction of u-PAR gene expression via the mitogen-activated protein kinase and c-Jun-NH2-kinase pathway (20, 26), and was required for the induction of gene expression brought about by mutation-activated K-ras (27). In another study, c-Jun binding to this motif was required for the activation of u-PAR gene expression by a constitutively active RalA involving a c-Src-intermediate (28). In addition, our previous studies in vitro suggested a functional synergism of this AP-1 motif with the AP-2/Sp1 promoter region 152/ 135, because cotransfection of GEO colon cancer cells with both an AP-2a and a JunD expression construct stimulated u-PAR promoter activity to an extent greater than the sum of the individual expression constructs (22). Taken together, these results suggested that the AP-1 region 190/ 171, besides the AP2/Sp1 region 152/ 135, is essential for mediating diverse means of u-PAR gene regulation in cultured colon cancer, in part being synergistic to the AP-2/ Sp1 motif. However, these studies were again based on cell lines and could not reflect the in vivo situation in resected tumors of patients. In particular, it did not answer the question as to what extent this motif might be specifically relevant in tumors in contrast to normal tissues. The aim of the present study was therefore (a) to elucidate the situation of AP-1-transcription factor binding to region 190/ 171 of the u-PAR promoter in a large series of resected tumor and corresponding normal tissues from patients with colorectal cancers, (b) to do a combinatorial analysis for transcription factor binding to both regions 190/ 171 and 152/ 135, compare tumor-specificity between both motifs, and potentially find supportive evidence from correlations in resected carcinomas for synergism, (c) to perform a first preliminary analysis of the potential clinical and prognostic effects of these transcription factor bindings. It is the first large-scale study giving evidence of the in vivo relevance of this AP-1-binding element of the u-PAR promoter in colorectal carcinoma, also providing data for a large subset of patients in which both promoter elements 190/ 171 and 152/ 135 were analyzed in parallel. The study suggests that AP-1 binding to region 190/ 171 is less tumor-specific than seen for the AP-2/Sp1 region. However, high percentages of transcription factor combinations binding to both promoter elements, the combination of AP-1 and the AP-2-like protein correlating with high u-PAR protein amounts, and a trend towards a poorer prognosis of patients showing binding of all three transcription factors, corroborate the hypothesis of synergism from data in resected tumors. The results indicate that such analyses can give a clear understanding for a careful subgroup selection of future clinical studies of an optimized tumor-selective targeting, involving pathways mediated by diverse promoter regions of the same gene.