| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Molecular Oncology, Markers, Clinical Correlates |
Department of Surgery, Klinikum Grosshadern, Ludwig Maximilians University of Munich, 81377 Munich, Germany [D. M. S., J. H. L., K. U. G., F. W. S., K. W. J., H. A.]; Department of Cancer Biology, M. D. Anderson Cancer Center, Houston, Texas 77030 [D. D. B., H. W.]; and Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, San Francisco, Comprehensive Cancer Center and CancerResearch Institute, San Francisco, California [E. R. L.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Experimental Design: In 145 gastrointestinal cancer patients, electrophoretic mobility shift analysis and supershift assays for u-PAR-promoter region -152/-135 were performed in tumors and corresponding normal tissues. u-PAR protein levels were measured by ELISA.
Results: Binding of Sp1 to region -152/-135 in tumors in contrast to corresponding normal mucosae was observed in 55% of colorectal and in 52% of gastric cancer patients. Tumor-specific binding of an AP-2-related factor was seen in 59% of colorectal and in 63% of gastric cancer patients. A significant correlation between AP-2 (P < 0.0001) and Sp1 (P = 0.0003) binding with a high u-PAR expression was observed in tumors but not in normal mucosae. Tissues of five nontumor patients did not show transcription factor binding to this region.
Conclusions: This is the first study to show the tumor-specific binding of trans-activators to the u-PAR promoter region (-152/-135) biochemically in a large series of resected tumors. For the subpopulation of 
60% of patients with tumor-restricted u-PAR-transactivation, a molecular targeting of this region or its activating pathways should be pursued as a new antimetastasis therapy approach.
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The urokinase receptor (u-PAR3 ) is a Mr 55,000–60,000 heavily glycosylated, disulfide-linked cell surface receptor that, by its ligand urokinase-type plasminogen activator, leads to a very efficient plasmin-mediated degradation of extracellular matrix components such as fibrin and collagen IV (1, 2, 3, 4, 5, 6, 7, 8) . Thus, it is one of the molecules promoting invasion and metastasis and has repeatedly been shown to be predictive of a poor clinical prognosis of diverse carcinomas such as breast and especially gastrointestinal cancers (9, 10, 11, 12, 13, 14, 15, 16, 17) . In cancer, a high expression of u-PAR is mainly brought about at the transcriptional level, although other mechanisms at the posttranscriptional level may be involved (18 , 19) . In our recent work using colon cancer cell lines (20) we had shown that, among other cis-elements, a motif spanning -152/-135 of the u-PAR promoter bound by an AP-2-like transcription factor and Sp1 was required for a high u-PAR gene expression. The AP-2-like protein mediated a high constitutive and phorbol 12-myristate 13-acetate-inducible expression. Likewise, Sp1 contributed to a high u-PAR gene expression caused by the c-src-oncogene, which is activated in >70% of colonic and gastric cancers (20, 21, 22) . With both a dominant negative AP-2 expression construct and a Src-inhibitor, u-PAR-mediated proteolysis in highly invasive colon cancer cell lines could be down-regulated considerably. This suggested that this promoter element mediating different pathways of u-PAR-induction could be a promising target to counter u-PAR-mediated metastasis in gastrointestinal cancers. However, these data were based on artificial cell lines and are far removed from the situation with naturally arising tumors.
For this study, we established a gelshift and supershift method that is able to give clear transcription factor binding results with resected tissue material. It is the first large study to provide evidence for the biological in vivo relevance of this AP-2/Sp1 promoter element in regulating u-PAR in a large series of resected colorectal cancers, as well as in a completely different biological tumor entity (gastric carcinomas), thereby postulating a broad regulatory principle. Our study is the first in which the clinical relevance of a transcriptional induction of gene expression for two different tumor types is demonstrated. The data suggest that in 60% of gastrointestinal cancer patients, the transactivation of u-PAR gene expression by AP-2/Sp1 is restricted to the tumor tissue, thus postulating a patient population in which this region could be used as a tumor-specific target.
|
PATIENTS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Prospective follow-up (physical exam, ultrasound, chest X-ray, tumor markers carcinoembryonal antigen, Ca 19-9 and Ca 72-4) was scheduled 6, 12, 18, 24 months after surgery and in 1-year intervals thereafter. Tumor recurrence was diagnosed by biopsy or explorative surgery, if possible. Causes of death were evaluated clinically. Of 5 patients operated for nonmalignant reasons, specimens could be taken for EMSA analysis as negative controls: 2 normal livers; 1 small intestine; 1 spleen; 1 smooth muscle; 1 peritoneum; 1 parathyroid gland; and 1 connective tissue. The study was approved by the institutional ethical standards committee and performed with the patients’ informed consent.
|
EMSA and Electrophoretic Mobility Supershift Analysis.
Nuclear extracts (25 µg) were incubated in a buffer [25 mM HEPES (pH 7.9), 0.5 mM EDTA, 0.5 mM DTT, 0.05M NaCl, 4% (v/v) glycerol] with 20,000 cpm of the [32P]ATP-end-labeled (phage T4 polynucleotide kinase, ICN catalogue no. 151935) oligonucleotide containing region -152/-135 of the u-PAR-promoter (5'-CCAGCCGGCCGCGCCCCGGGAAGGGA-3') for 10 min in the absence or presence of a 200-fold excess of unlabelled oligonucleotide to show specificity of the binding (room temperature). A total of 0.5 µg of poly(deoxyinosinic-deoxycytidylic acid) was present in each reaction to block unspecific binding. The reactions were subjected to gel electrophoresis on a 5% polyacrylamide gel containing 5% glycerol and exposed to Kodak MR film for 24 h (-80°C, intensifying screen). For supershift analysis, 1 µg of antibody (anti-Sp1, anti-Sp2, anti-Sp3, anti-Sp4, and rabbit IgG control, Santa Cruz Biotechnology; catalogue nos. sc-59, sc-643, sc-644, sc-645, sc-2338, respectively) was added to the reactions 10 min after nuclear extracts and oligonucleotide had been incubated. Supershift reactions were incubated for 60 min at 4°C and electrophoresed at 4°C to ensure complex stability. The binding of the AP-2-related factor was competed with an AP-2
consensus oligonucleotide with the sequence 5'-GATCGAACTGACCGCCCGCGGCCCGT-3' (Santa Cruz Biotechnology; catalogue no. sc-2513).
As negative controls, a lane without nuclear extract and a second lane without poly(deoxyinosinic-deoxycytidylic acid) and nuclear extract were processed in each EMSA.
Quantification of EMSA Results for Statistical Analysis with u-PAR-ELISA.
Binding activities of Sp1 and the AP-2-related factor were quantified by densitometry using a high resolution image scanning system (Hewlett Packard) and the evaluation software ScanPack 3.0 (Whatman-Biometra, Goettingen, Germany), which determined the area under the densitometric curve in each tissue sample for the Sp1 and AP-2 bands. Binding activities were expressed as a ratio between the densitometric binding activities of Sp1, Sp3, and the AP-2-like protein in the tissues and activities from equal amounts of a standardized RKO nuclear extract, which was processed in each EMSA in parallel as a reference. The RKO binding activity was thereby regarded as 100%.
Qualitative presence of transcription factor binding was defined as soon as a band could be detected by densitometry above the background of the shift. The term high binding was defined for a transcription factor as soon as the binding intensity of its corresponding band was ≥40-fold density of the background of the individual shift. In Fig. 1
, Lane 19, the Sp1 band gives the example of the lowest intensity in a tumor tissue that was defined as high binding in our tumors, the intensity of this band corresponding to 40-fold of the background. In the same tumor (Fig. 1
, Lane 19), the AP-2-like band can be seen as well, however, was not deemed as high binding because its intensity was below this density limit. In contrast, the term low binding was defined by the intensity of the transcription factor band ≤ 10-fold of the background of the individual shift, a density corresponding to a band just visible as a specific band to the human eye. In Fig. 1
, the normal tissue in Lane 5 gives the example for the highest density of a band (Sp1) that was considered low binding, the Sp1 band in this example corresponding to 10-fold of the individual background of the experiment. The term tumor-specific binding was attributed to a case in which a high transcription factor binding according to the definition above was seen in the tumor in parallel to an absent or low binding according to the definition above in the corresponding normal tissue. As an example, the binding in patient 4 (Fig. 1)
was tumor specific for Sp1 but not for AP-2. All other patients in Fig. 1
show a tumor-specific binding for both transcription factors.
|
Western Blotting.
Western blotting was performed as described previously (20
, 21) . Briefly, tissue extracts prepared according to the protocol above were subjected to standard Western blotting. The blot was probed with a polyclonal anti-Sp1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA; catalogue no. sc-59X, 1:500) detecting the Mr 95,000 native and the Mr 106,000 phosphorylated form of Sp1 (23)
and a horseradish peroxidase-conjugated antirabbit IgG (Amersham, Freiburg, Germany; 1:10,000). Bands were visualized by enhanced chemiluminescence (Amersham). Direct reprobing as an internal control was performed using an anti-ß actin antibody (Sigma, St. Louis, MO; catalogue no. A5441, 1:5000).
Statistical Analysis.
Statistical analysis was performed using StatView 5.0 for Macintosh/Windows. The correlation between u-PAR amounts (measured quantitatively with ELISA) and transcription factor binding in tissue samples (densitometry, see above) was determined by linear regression analysis. For correlation analysis between the presence of transcription factor binding and established tumor classifications, transcription factor binding was dichotomized (presence/absence, see above), and 
2 tests (Bonferroni corrected) were performed, considering the parameters as follows: Dukes’ stage as A–D; Laurén’s classification as intestinal versus diffuse/mixed; lymphangiosis carcinomatosa as presence versus absence; pT as pT1–2 versus pT3–4; pN as pN0 versus pN1–3; and M, Unio Internationale Contra Cancrum, G, and Borrmann stages according to the established groups. Differences between tumor and normal tissue in u-PAR-content were calculated with the Wilcoxon test for independent values. P < 0.05 was defined as significant, a P < 0.1 as a trend.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
with primary colorectal (n = 118) or gastric cancer (n = 27) were analyzed for transcription factor binding in their primary tumors and corresponding normal mucosae. A series of 17 colorectal cancer patients having been tumor-resected at the M. D. Anderson Cancer Center Houston (13 of them R0) were analyzed first as a preliminary series. The confirmatory prospective series of 128 patients operated at the Ludwig Maximilians, University of Munich (96 of them R0), was screened independently. The median time of follow-up was 39 months (range, 2–95 months) for the preliminary series and is presently 22 months (range, 2–42 months) for the confirmatory series. During this time, 7 and 13 patients died, respectively. In the curatively resected (R0) patients, 5 and up to now 10 recurrences occurred in the preliminary and in the confirmatory series, respectively.
Qualitative Results of EMSA and Electrophoretic Mobility Supershift Analyses.
All EMSA analyses were performed using a 32P-end-labeled oligonucleotide corresponding to region -152/-135 of the u-PAR-promoter and normalized equal protein amounts of normal and tumor tissue extracts, as well as equal amounts of a standardized nuclear extract of RKO colon cancer cells as a reference. In previous publications, we had shown that binding of an AP-2-like transcription factor, Sp1, and also Sp3 to region -152/-135 of the u-PAR-promoter is reproducibly seen with this high u-PAR-expressing colon cancer cell line, represented by three specific bands of slower mobility (Figs. 1
and 2
, Lane 3; Ref. 20
). The upper of these bands represents Sp1 binding, the intermediate Sp3 binding, and the lower the binding of the AP-2-like transcription factor as demonstrated previously (20)
.
|
, Lane 24). Sp3 binding in tissues was investigated using an antibody directed against the DNA binding domain of Sp3, resulting in a reduction of the corresponding slower migrating band in some tissues (Fig. 2, A and B
, Lanes 19, #). The binding of Sp1 was indicated by the ability of an antibody against Sp1 to supershift the slowest migrating band (Fig. 2A
, Lanes 5, 11, 17, and 23, and Fig. 2B
, Lane 17, *). Transcription factor binding in RKO cells and in the tissues was specific as a 200-fold excess of the nonlabeled oligonucleotide corresponding to region -152/-135 of the promoter resulted in an elimination of the bands (Fig. 1
, Lanes 4, 6, 8, 10, 12, 14, 16, 18, and 20, indicated by +).
To gain additional information on as to whether an increase of transcription factor binding correlates with increased expression, we performed Western blotting analysis for Sp1 for 10 selected tumors and the corresponding normal tissues. Examples of this protein analysis are given in Fig. 3
. As can be seen from these, there is not necessarily an up-regulation of Sp1 protein amounts in tumor tissues in contrast to normal tissues when high transcription factor binding is observed in the former. In examples N8 and TU8, Sp1 expression is several fold higher in the normal than in the tumor tissue (Fig. 3)
, but the binding of this transcription factor to region -152/-135 of the u-PAR promoter is hardly visible in the normal tissue, this in contrast to a high binding in the tumor tissue. Statistically, there was no significant correlation between Sp1 binding (EMSA) and protein amounts (Western) in these tissues. These data emphasize that it is essential to screen for transcription factor binding to a specific promoter element rather than for mere protein levels, which cannot give information on transcription factor function.
|
. In the preliminary series, high binding of the AP-2-like protein to region -152/-135 of the u-PAR-promoter was detected in the tumor tissue in 70.6% and high binding of Sp1 in 41.2% of the cases, this in contrast to an absence of binding in the normal mucosae.
|
. More importantly, a high transcription factor binding in the resected tumors in contrast to a low or absent binding in the correspondingly investigated normal mucosae (tumor-specific binding) was observed for the AP-2-like factor in 57.4% of colorectal and 63.0% of gastric cancer patients. For Sp1, a tumor-specific binding to region -152/-135 of the u-PAR-promoter was detected in 58.4% of colorectal and 51.9% of gastric cancer patients. In addition, an analysis of diverse normal tissues of 5 patients operated for nonmalignant reasons (2 livers, 1 small intestine, 1 spleen, 1 connective tissue, 1 smooth muscle, 1 peritoneum, and 1 parathyroid gland) revealed no transcription factor binding to this u-PAR promoter region. These data indicate that the potential transactivation of the invasion-related gene u-PAR by the AP-2-like protein or Sp1 is tumor-specific in more than one-half of the patients with colorectal or gastric cancer.
Tumor-specific binding of both the AP-2-like protein and Sp1 together was observed in 51.5% of colorectal and 37.0% of gastric carcinoma patients (Table 2)
in the prospective series, implicating that both pathways may be active in these tumors and opening the possibility of a parallel targeting of both mechanisms. The presence of binding of the AP-2-like transcription factor and Sp1 in the tumors correlated significantly (P < 0.001, 2).
Tumor-specific binding of Sp3 to region -152/-135 in the prospective series was observed in 56.4% of colorectal and 48.1% of gastric cancers, respectively (Table 2)
.
In some of the normal tissues (Fig. 1
, Lanes 5, 13, and 17), transcription factor binding was also detectable as compared with the tumors. This might be because of multiple functions of the factors analyzed (e.g., regulation of housekeeping genes by Sp1), leading to a still remote binding activity for the u-PAR cis-element. In order for a patient to fulfill the definition of tumor-specific binding, however, the binding in the normal tissue had to meet the criteria for low or absent binding defined in the "Patients and Methods" section.
Correlation of Transcription Factor Binding to Region -152/-135 of the u-PAR-Promoter with Established Tumor Characteristics.
Next, potential correlations between the binding of transcription factors to region -152/-135 of the u-PAR-promoter and established tumor parameters were analyzed (2 analysis, Bonferroni corrected). In the prospective confirmatory series, there was a significant correlation of the binding of the AP-2-like protein in colorectal tumor tissues with M stage (distant metastasis, P < 0.001) and a trend of a correlation between Sp1 in the gastric tumor tissues and lymphangiosis carcinomatosa (P = 0.091). In the preliminary series, a trend was observed for Sp1 binding and a positive pN stage (lymph node metastasis, P = 0.094).
Correlation of Transcription Factor Binding with u-PAR Protein Amounts in Tumor and Normal Tissue.
To corroborate the biological relevance of the transactivation of u-PAR gene expression by region -152/-135 in tumor versus corresponding normal tissue, correlations between EMSA results and u-PAR-protein amounts as measured by ELISA were investigated. EMSA results of the resected tissues were quantified by densitometric scanning of the bands of slower mobility corresponding to transcription factor binding in relation to the densitometric intensity of the bands, resulting from equal amounts of a standardized RKO-nuclear extract, which was processed in parallel in each EMSA. This method was applied by others (24)
for a relative quantitation of EMSA data.
Median u-PAR protein amounts were 1.6 ng/mg protein (range, 0.4–7.9 ng/mg) in tumor tissues and 0.7 ng/mg (range, 0.2–1.3 ng/mg) in normal tissues in the preliminary series, 1.0 ng/mg (range, 0.0–4.6 ng/mg) in tumor tissues, and 0.2 ng/mg (range, 0.0–1.2 ng/mg) in normal tissues in the confirmatory series, respectively. The differences between tumor and normal tissue in u-PAR content were significant with P = 0.0045 and P < 0.0001, respectively (Wilcoxon). In linear regression analysis (Table 3)
, a significant correlation between the binding of the AP-2-like factor (P = 0.0025) with high u-PAR protein amounts was found in the carcinoma tissues in the preliminary series. Also, in the independent confirmatory series, a significant correlation between AP-2-like binding (P < 0.0001) and also Sp1-binding (P = 0.0003) with high u-PAR protein amounts in the tumors was detected. This was especially attributable to the highly significant association in the 101 colorectal carcinomas of this series (both P < 0.001). The binding of Sp3 in tumor tissue also correlated significantly with high u-PAR protein amounts (P = 0.035 and P = 0.027, respectively; Table 3
).
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
60%) in whom a tumor-specific transactivation of the invasion-related gene u-PAR via promoter region -152/-135 is evident and predicted from tumor-specific transcription factor binding in these cases. The data additionally suggest that the transactivation of u-PAR gene expression in two biologically different tumor entities of the gastrointestinal tract (colorectal and gastric) may be of biological relevance in tumor tissue only, as can be hypothesized from a significant correlation of transcription factor binding with a high u-PAR expression exclusively in the tumor tissue. Therefore, we suggest the activation of u-PAR gene expression in gastrointestinal cancers via promoter region -152/-135 as a first transcriptional mechanism that might have the potential of clinical and therapeutic relevance, combined with the possibility to screen for appropriate patients with the tissue gelshift method. The only existing study concerning this issue thus far (24)
analyzed Sp1 binding to an artificial Sp1 consensus oligonucleotide (Promega) in 14 breast carcinomas as compared with 5 benign breast lesions and found a significant correlation between Sp1 binding to this artificial consensus motif with elevated u-PAR expression in contrast to the benign breast lesions. Nevertheless, an additionally larger study using the natural u-PAR gene promoter to our knowledge has not been conducted thus far, nor has a supershifting method been established thus far showing a convincing identification of Sp1 in a large series of resected tissues, as it can be seen in our series.
The u-PAR has repeatedly been shown to promote invasion and metastasis in gastrointestinal cancers, and its overexpression is associated with a poor patient prognosis (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
. In our previous work using colon cancer cell lines, we had shown that an AP-2-like transcription factor bound to region -152/-135 of the u-PAR promoter mediated a constitutively high and phorbol 12-myristate 13-acetate-inducible u-PAR expression in a highly invasive colon cancer cell line (20)
. This AP-2-like protein found in our previous work in cell lines and also in our tumor tissues presented here may be a new isoform of an AP-2-like factor because it cannot be supershifted with an AP-2-antibody in contrast to authentic AP-2 at promoter motif -152/-135 but is eliminated by immunoprecipitation of the nuclear extract with an AP-2-antibody (20)
. Furthermore, it shows a different mobility than authentic AP-2 protein in gelshift experiments using our -152/-135 oligonucleotide but can be competed with an AP-2 consensus motif (20)
. As a supershift is not possible, this new AP-2 isoform would potentially also not be detected by known antibodies against AP-2 subtypes in Western blotting so that the gelshift method represents the only way of detection of this isoform up to now. The results presented here showing a large subgroup of patients with a tumor-specific binding of this AP-2-like protein is promising because this might implicate an AP-2 isoform that gets activated in tumor tissue in contrast to normal tissue.
The binding of Sp1 to promoter region -152/-135, on the other hand, was crucial for the induction of u-PAR expression by the c-src-oncogene in cultured colon cancer in previous studies (21) . u-PAR-mediated proteolysis could be inhibited by both transfection with a dominant negative AP-2 and a specific Src inhibitor, which would propose both the AP-2-like- and the Sp1-regulated u-PAR expression as attractive targets for new cancer therapy concepts. However, to support this notion, it was essential to show that the transactivation of u-PAR gene expression by these factors is not an artifact of cell lines but rather observed in human resected tumors.
In this study, we demonstrate that, taking together the preliminary and the independent confirmatory series (the slight differences between the two series being possibly explained by different patient numbers or population-dependent differences), in 60% of cases, a tumor-specific binding of the AP-2-like transcription factor and/or Sp1 to region -152/-135 of the u-PAR promoter is observed, whereas in equal amounts of nuclear extracts of corresponding normal mucosae the binding of these factors is almost absent. This was shown not only for colorectal but also for gastric cancer, which are tumors of different biological entity. In addition, other normal tissues as liver, small intestine, and spleen did not show transcription factor binding to this u-PAR promoter region. Moreover, a significant correlation between AP-2 and Sp1 (and also Sp3) binding and a high amount of endogenous u-PAR was observed in the tumor tissues only but not in normal tissues. This proposes that an activation of u-PAR gene transcription by these factors bound to region -152/-135 is biologically relevant in the tumor tissues only and that these transactivating pathways potentially get activated during malignant progression. The fact that a high expression of Sp1 at the protein level in Western blotting analysis of selected samples was not necessarily associated with a high binding of this factor to promoter region -152/-135 emphasizes the necessity to perform gelshifting rather then mere Western blotting to get information on transcription factor function at this specific cis-element. This notion is supported by the fact that for transcription factors as Sp1 or Sp3, it is known that their ability to transactivate at target promoter elements is not increased by an induction of Sp1/Sp3 synthesis and protein amounts but rather by posttranslational modifications such as phosphorylation and/or glycosylation, which, in turn, modifies their transactivating potential or DNA binding affinity (25)
.
In some normal tissues, there is still a weak transcription factor binding, which could be explained by the fact that, especially for ubiquitous transcription factors such as Sp1, which are needed to transactivate housekeeping genes in most tissues, these transcription factors are present but have a very low affinity to the u-PAR-promoter element investigated. This would create a background binding that is not relevant for u-PAR regulation in vivo.
Certainly, the data cannot prove a functional role of transcription factor binding in the resected tissue as it can be done with, for example, promoter reporter assays in cell lines. To our knowledge, however, it is technically not yet possible to perform functional reporter assays such as luciferase or chloramphenicol acetyl transferase assays in authentic resected tissues, and our strategy applying tissue gelshifting and comparisons with u-PAR expression presented in this article might be one of the closest technically feasible approximations thus far. Nevertheless, taking together our previous functional results (20 , 21) and the results on resected tissue presented here, a functional and clinical importance of especially AP-2-type and Sp1 transcription factor binding to region -152/-135 of the u-PAR promoter in tumor-associated proteolysis and metastasis is very likely. Especially, the possibility to identify the subpopulation of patients in whom a tumor-specific transactivation can be postulated quite easily with our tissue gelshift assay suggests a good applicability for the clinic. It is certainly possible to broaden this protocol to any other promoter cis-element of interest regulating a gene relevant for malignant transformation or tumor progression from which a preferentially transcriptional induction of gene expression is known and to integrate the assay as a quickly obtainable parameter in new molecular staging concepts. In fact, our results indicating a subpopulation with tumor-specific binding emphasize the necessity of performing a screening of patients before a new therapy is applied; as, in our case, for the 40% of patients that do not show tumor-specific binding, a future targeting of this region would not be tumor specific and potentially would have side effects on normal mucosa. Therefore, in our opinion, it will be essential in the near future to broaden established pTNM staging models for tumors by relevant functional molecular factors (for example, the method shown in this article but many more others) and to be able to characterize the subpopulations of patients that can safely and successfully be treated with new molecular targeting strategies.
Our data are corroborated by first clinical results demonstrating that a nuclear localization of AP-2-type transcription factors correlated with an unfavorable clinical outcome in prostate and ovarian cancer (26
, 27)
. Also, it has been demonstrated that the DNA-binding activity in addition to an increased expression of AP-2 was evident in transformed but not in normal human fibroblasts (28)
. Furthermore, it is known that other invasion-related genes such as MMP-2, MMP-9, cathepsin D, cathepsin B, and MT1-MMP are regulated by AP-2-like transcription factors or/and Sp1 (in part, also Sp3), this in part via combined AP-2/Sp1/(Sp3)-promoter motifs (28, 29, 30, 31, 32)
, potentially implicating a differential means of invasion/metastasis regulation by these parameters.
The exact functional role of Sp3 binding to region -152/-135 of the u-PAR-promoter is not clear yet. Other reports on combined AP-2/Sp1/Sp3 motifs have shown that it can act as an enhancer or inhibitor of AP-2/Sp1 transactivation in different genes, among them, interestingly, the c-src gene (29 , 33) . Our clinical results in this study, which implicate a significant positive correlation of Sp3 binding with u-PAR gene expression in the tumor tissue, suggest that it might act as an in vivo activator at this specific motif. However, in some tissues we investigated, the band corresponding to Sp3 binding in the RKO control extract was not clearly reduced when the Sp3 antibody was used in the resected tissue, so it could be hypothesized that this band in some of the tissues investigated might represent the binding of a transcription factor different from Sp3 with up to now unknown identity. This is very interesting because it would mean a difference to our previous results seen in cell lines (20) and show the importance of corroborating cell line results by ex vivo tumor tissues when dealing with molecular biology results. Therefore, it needs to be emphasized that a definite speculation on the role of Sp3 in resected tumor tissues cannot be done by the present work, as opposed to the conclusions drawn for Sp1 and the AP-2-like factor.
Nevertheless, taken together, our data suggest promoter region -152/-135 of the invasion-related gene u-PAR as a potential target for a relatively tumor-specific approach in developing new molecular antimetastatic strategies. Currently, translational studies evaluating the in vivo applicability of a specific transcriptional targeting strategy involving the AP-2/Sp1 transactivation via promoter region -152/-135 and involving small molecular Src-inhibitors (34) whose effect is mediated via region -152/-135 are under way in our group.
|
FOOTNOTES |
|---|
1 H. A. is supported by the Dr. Mildred Scheel Stiftung (Deutsche Krebshilfe), Bonn, the Wilhelm Sander Stiftung, Munich, Germany, the Faculty of Medicine of the Ludwig Maximilians University of Munich, the Friedrich-Baur-Stiftung, the Graute-Oppermann-Stiftung, and the Muenchener Medizinische Wochenschrift Board of Editors, Munich, Germany. This publication contains parts of the dissertation of Denis M. Schewe performed in partial fulfillment of the requirements for the Dr. medicinae at the Faculty of Medicine, Ludwig Maximilians University, Munich, Germany. 
2 To whom requests for reprints should be addressed, at Molecular Oncology, Department of Surgery, Klinikum Grosshadern Ludwig Maximilians, University of Munich, 81377 Munich, Germany. Phone: 49-89-7095-1; Fax: 49-89-7004418; E-mail: DAllgayer{at}aol.com
3 The abbreviations used are: u-PAR, urokinase-type plasminogen activator receptor; AP, activator protein; MMP, matrix metalloproteinase; EMSA, electrophoretic mobility shift assay.
Received 5/23/02; revised 1/30/03; accepted 2/19/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-related factor. J. Biol. Chem., 274: 4702-4714, 1999. predicts survival in epithelial ovarian cancer. Br. J. Cancer, 82: 1974-1983, 2000.[CrossRef][Medline]
, and can be inhibited by triplex-forming oligonucleotides. J. Biol. Chem., 275: 847-854, 2000.This article has been cited by other articles:
|
|
 |
|
  G. D. Maurer, J. H. Leupold, D. M. Schewe, T. Biller, R. E. Kates, H.-M. Hornung, U. Lau-Werner, S. Post, and H. Allgayer Analysis of Specific Transcriptional Regulators as Early Predictors of Independent Prognostic Relevance in Resected Colorectal Cancer Clin. Cancer Res., February 15, 2007; 13(4): 1123 - 1132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Schewe, T. Biller, G. Maurer, I. A. Asangani, J. H. Leupold, E. R. Lengyel, S. Post, and H. Allgayer Combination Analysis of Activator Protein-1 Family Members, Sp1 and an Activator Protein-2{alpha}-Related Factor Binding to Different Regions of the Urokinase Receptor Gene in Resected Colorectal Cancers Clin. Cancer Res., December 15, 2005; 11(24): 8538 - 8548. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Wang, L. Yang, Md. S. Jamaluddin, and D. D. Boyd The Kruppel-like KLF4 Transcription Factor, a Novel Regulator of Urokinase Receptor Expression, Drives Synthesis of This Binding Site in Colonic Crypt Luminal Surface Epithelial Cells J. Biol. Chem., May 21, 2004; 279(21): 22674 - 22683. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Trisciuoglio, A. Iervolino, A. Candiloro, G. Fibbi, M. Fanciulli, U. Zangemeister-Wittke, G. Zupi, and D. Del Bufalo bcl-2 Induction of Urokinase Plasminogen Activator Receptor Expression in Human Cancer Cells through Sp1 Activation: INVOLVEMENT OF ERK1/ERK2 ACTIVITY J. Biol. Chem., February 20, 2004; 279(8): 6737 - 6745. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
