Increased expression of transcription factor TFAP2α correlates with chemosensitivity in advanced bladder cancer

Bladder cancer is the fifth most common malignancy in Europe and the fourth most common malignancy in the United States. The most prevalent histological type is transitional cell carcinoma (TCC), which constitutes up to 95% of the malignancies of the bladder. About 30% of TCC's display a solid, invasive growth pattern, being either locally advanced (pT3a, pT3b, pT4 and/or pN1, pN2, pN3 M0) or metastatic (M₁) at the time of diagnosis or at later visits [1]. The response rate to chemotherapy is only approximately 50% [2]. Presently, there are two standard chemotherapeutic regimens: MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) or GC (gemcitabine and cisplatin). Median survival is 14 to 15 months, and 5-year overall survival rate is between 13% and 15% [3]. Although the gemcitabine and cisplatin combination has a significantly better toxicity profile, both regimens still carries risk for significant toxicity and toxic deaths [4] and a substantial fraction of patients will suffer from adverse reactions without achieving any benefit. It is therefore of utmost importance to be able to discriminate between responders and non-responders for improved selection of patients to chemotherapy and to improve the individual patient's quality of life. A molecular signature for chemo-resistance, based on microarray profiling of cancer specimens has been described for locally advanced and/or metastatic bladder cancer [5]. BIRC5 (survivin) and BSG (emmprin) were validated as independent predictive markers for response and survival after cisplatin-containing chemotherapy by immunohistochemistry in an independent material of 124 patients with locally advanced (T_4b and N_2-3) or metastatic (M₁) disease [5]. In the present study we investigated another interesting molecule from this gene expression signature, the transcription factor Activator Protein TFAP2α. TFAP2α belongs to the TFAP2 family of transcription factors that in humans and mice consist of five members, TFAP2α, TFAP2β, TFAP2γ, TFAP2δ and TFAP2ε. Orthologs show a similarity between 60 and 99% at the amino-acid level. The TFAP2 family is characterized by having a highly conserved helix-span-helix dimerization motif at the carboxyl terminus together with a central basic region and a less conserved proline/glutamine rich domain at the amino terminus. The helix-span-helix domain facilitates homo and heterodimerization between the TFAP2 members. Once dimerized, the helix-span-helix motif and the neighboring basic domain facilitate DNA binding and the N-terminal proline/glutamine-rich domain mediates transactivation. The TFAP2 family members participate in the regulation of many signaling pathways and are essential during embryogenesis and development. TFAP2 proteins participate in tumorigenesis through regulation of neoplasia associated genes such as P21, Rb, TP53, ERα, BCL2, cKIT, MMP-2, E-cadherin and c-myc (reviewed in [6, 7]). TFAP2α knockout mice are not viable and have severe ventral body wall closure defects (thoracoabdominoschisis)[8]. Currently many studies have linked deregulated TFAP2α activity to malignant transformation. The TFAP2α gene locus at 6p22 is frequently lost in various cancers [9]. Lost or decreased TFAP2α expression has been identified in human cancers of the breast, colon, prostate, ovary and brain [10‐15] suggesting TFAP2α to be a tumor suppressor gene. Immunohistochemical analysis demonstrated a correlation between decreased TFAP2α expression and advanced colon adenocarcinomas [11]. In breast cancer, low nuclear TFAP2α expression is associated with disease progression and elevated metastatic capability [12]. Furthermore, reduced TFAP2α expression predicted elevated risk of recurrent disease in breast cancer [16]. Moreover, re-expression of TFAP2α in metastatic melanoma cells decreased their tumorigenity and inhibited their metastatic potential in nude mice [17]. So far the importance of TFAP2α in transitional cell carcinomas of the bladder has not been described. In the present study, we showed that high levels of nuclear and cytoplasmic TFAP2α protein were associated with increased overall survival and progression free survival of patients with lymph node positive advanced bladder cancer after cisplatin based chemotherapy. Furthermore, we showed that siRNA mediated knockdown of TFAP2α stimulated proliferation of the SW780 bladder cell line along with decreased cisplatin and gemcitabine sensitivity.

Archival formalin-fixed paraffin embedded (FFPE) tissues of cystectomy specimens were derived from a previously reported Phase III, multicenter randomized control trial of two different adjuvant chemotherapy regimens, AUO-AB 05/95 conducted by the Arbeitsgemeinschaft Urologische Onkologien (AUO) collaborative group of the Federal Republic of Germany [18, 19]. The clinicopathological characteristics of all patients were reviewed by one surgical pathologist (A.H.). Tumor staging was performed according to the criteria of the International Union against Cancer (UICC). Representative haematoxylin- and eosin-stained (HE) slides of formalin-fixed and paraffin-embedded tissue blocks obtained at cystectomy where reviewed and the tumor area was marked. Tissue microarrays (TMA) where constructed by obtaining a 1.5 mm punch-biopsy from each tumor. TMA construction was done as described previously [20]. The TMAs were stained with HE and were reviewed for the presence of representative tumor tissue by one surgical pathologist. The study was approved by the local ethical committee of the University Regensburg.

Studies investigating TFAP2α expression in human cancers show a correlation between reduced nuclear expression of TFAP2α and shorter recurrence-free survival and aggressive clinicopathological features in colon cancer [11], breast cancer [16], ovarian cancer [13] and melanoma [24]. Several in vitro studies demonstrated that TFAP2α has an inhibitory effect on cell proliferation and tumorigenesis and proposed that TFAP2α has a tumor suppressive effect in malignancies, although articles with opposing conclusions also have been published [7]. A large breast cancer study with immunohistochemical detection of TFAP2α revealed that a reduced level of TFAP2α in the nucleus and/or a shift of the protein to the cytoplasm may predict a shorter recurrence-free survival and breast cancer-related survival [7]. The observed correlation between high nuclear TFAP2α staining and decreased overall survival rate for the non lymph node invasive subgroup is in agreement with a breast cancer study from Finland that demonstrated that particular for lymph node positive patients low nuclear TFAP2α expression is associated with disease progression and elevated metastatic capability [25]. Reduced TFAP2α expression also predicted elevated risk of recurrent disease in breast cancer [16]. A study of human melanomas showed that high level of TFAP2α expression in the cytoplasm relative to the nucleus correlates with poor prognosis. The loss of nuclear TFAP2α expression was associated with malignant transformation and progression of melanoma, indicating that translocation of TFAP2α from the cytoplasm to the nucleus may be disrupted during melanoma progression. Thus it seems that the major deficiency in TFAP2α activity in metastatic melanoma is the loss of nuclear translocation. This could be due to modifications in the nuclear-pore complexes or in the activity of the transport receptors (karyopherines/importins/exportins) or changes of the TFAP2α protein itself. In our study, both low nuclear and cytoplasmic expression predicted poor outcome, suggesting that decrease of TFAP2α transcription/translation or increased turnover rate are a more likely course than translocation in the case of bladder cancer. For the group of patients without lymph node invasion the association was shifted, high nuclear staining was associated with decreased survival time. Although the correlation is not highly significant the difference from the lymph node invasive group is striking. There is no other study analyzing the TFAP2α staining in an isolated group of patients without lymph node invasion. The results of the TFAP2α staining analysis raised the question if the decreased chemo sensitivity of bladder cancer patients with lymph node invasion having low TFAP2α staining is due to depletion of TFAP2α or deregulation of a downstream target. We performed knock down experiments in bladder cell lines and subsequently measured their sensitivity against cisplatin and gemcitabine. Previous studies have shown that the tumor suppressor activity of TFAP2α is mediated through a direct interaction with TP53. Furthermore, TFAP2α induces TP53 dependent p21 transcriptional activation. This supports the observed ability of TFAP2α to induce G1 and G2 cell cycle arrest in TP53 ^+/+ but not in TP53 ^-/- HCT116 colon cells [23]. In contrast to this, a study with breast cancer cell lines demonstrated that TFAP2α down regulation decreases chemosensitivity irrespective of their TP53 status [26]. In light of these observations, we decided to use a TP53 mutated and a TP53 wild type bladder cell line to conduct functional chemosensitivity studies. We chose the T24 (TP53 homozygous Y126*) and the SW780 TP53 wild type line, both having approximately the same dose response profile of cisplatin and gemcitabine making them appropriate to compare. We showed that TFAP2α silencing rendered SW780 less sensitive against cisplatin and gemcitabine induced cell death and potentiated the cell death of T24 cells. Moreover, we found that TFAP2α down regulation stimulated proliferation of SW780 cell and did not change the proliferation rate of T24 cells. The SW780 cell line metastasizes to regional lymph nodes in nude mice tumor transplants, corresponding to the clinical findings in lymph node positive patients [27, 28]. In contrast, TFAP2α silencing augments cisplatin and gemcitabine sensitivity and did not stimulate proliferation in the TP53 mutated and non-tumorigenic T24 bladder cell line, corresponding to clinical findings in lymph node negative patients. As mammalian cell terminal differentiate they undergo cell cycle arrest exiting from the cell cycle. TFAP2α mediates its role as a differentiation associated transcription factor through positive regulation of p21 thereby negatively regulating the cell cycle. TFAP2α induces expression of p21. The p21 promoter contains a TFAP2α binding site located at -103 and -95 where TFAP2α binds directly and stimulate expression [29]. In addition TFAP2α targets the p21 promoter in the p53 binding region at -2250, however only in the presence of p53 in agreement with TFAP2α has been shown to bind P53 in vivo and in vitro [23, 30]. Furthermore TFAP2α induces p21 dependent P53 expression corroborating the observed ability of TFAP2α to induce G1 and G2 cell cycle arrest [23, 31]. The explanation for the aberrant chemosensitivity of TFAP2α silenced T24 and SW780 could therefore be due to deregulation of the p53/p21 pathway in T24. For SW780, TFAP2α knockdown may suppress p53/p21 activation because TFAP2α is a positive regulate of p53/p21 expression. This potentiates the effect of cisplatin and gemcitabine as well as relief part of the suppression mediated by p53/p21 on the cell cycle. Taken together, our findings in T24 and SW780 cells may suggest that TFAP2α down regulation in bladder cells decreased cisplatin and gemcitabine sensitivity in a p53/p21 dependent manner. This is in line with studies showing that knock down of the TFAP2α expression in breast cancer and colon cancer cell lines resulted in significant reduction in chemotherapy-induced apoptosis [26, 32]. In non-small cell lung cancer, expression of p53 and p21(Waf1) in mediastinal lymph node specimens were significantly related to the response to platinum chemotherapy [33]. Moreover, overexpression of TFAP2α expression in a breast cancer cell line augmented increased chemosensitivity and induced endogenous TFAP2α protein levels in a posttranscriptional way [26]. Within our study, re-introduction of TFAP2α in T24 and SW780 was performed by transient and stable transfection, Transient transfection was very low (<20%) as monitored by QPCR/WB and the selected clones seem to loose TFAP2α as no increased in transcript was measured (QPCR) compared to the mock transfected. In the literature, re-introduction of TFAP2α into TFAP2α-negative SW480 colon cancer cells stimulates expression of E-cadherin and down regulation of the MMP-9 expression and leads to dramatic loss of cellular invasive potential in vitro. Interestingly, node positive colorectal cancers showed significant losses for p21 and E-cadherin compared to node negative cancer [34]. TFAP2α directly binds to the promoter of E-cadherin, where it has been previously reported to act as a transcriptional activator [35]. High E-cadherin expression has been reported to increase cisplatin and gemcitabine sensitivity in pancreatic cancer [36]. Furthermore stable transfectants expressing TFAP2α in the SW480 cell line significantly inhibited their growth in an orthotopic animal model [35]. Previous studies have also demonstrated that re-expression of TFAP2α in SW480 cells resulted in an inhibition of colony formation in vitro and upregulation of p21 ^Waf1/Cip1 [31].

In this study we demonstrated that high levels of nuclear and cytoplasmic TFAP2α protein was a predictor of improved overall survival and progression free survival of patients with locally advanced bladder cancer undergoing cisplatin based chemotherapy treatment when focusing on the lymph node invasive subgroup. In contrast, high nuclear TFAP2α staining was associated with decreased overall survival rate for the patients without lymph node metastases. Moreover, siRNA directed knock down of TFAP2α stimulated proliferation of SW780 bladder cells and decreased their cisplatin and gemcitabine sensitivity. On the contrary, TFAP2α silencing potentiated cisplatin and gemcitabine sensitivity and did not stimulate proliferation in the TP53 mutated T24 bladder cell line. Future studies are needed to further validate the predictive potential of TFAP2α expression in bladder cancer. Currently, we are collecting bladder tumor samples (locally advanced T3-4, N1-3 and/or metastatic M1) from patients that have been treated with cisplatin based chemotherapy and which have been characterized according to the RECIST response criteria. This cohort will be used to evaluate if TFAP2α staining and expression are predictive for cisplatin response in addition to survival.