Survivin is a member of the IAP (inhibitor of apoptosis) gene family and is markedly overexpressed exclusively in cancers but not in normal adult tissues [1-7] implying that it could be an ideal target for cancer directed therapy. Strong expression of survivin in tumors correlates with a poor cancer treatment response and poor outcomes. Notably, survivin is also present in some nonmalignant cells such as myeloid stem cells and peripheral blood mononuclear cells, T lymphocytes [6], melanocytes [8] and in hyperplastic polyps and sessile serrated adenomas [9]. Furthermore, survivin is widely expressed during embryogenesis in many tissues like human fetal lungs, liver, heart, kidney and gastrointestinal tract [5]. Apart from its role in antiapoptosis, survivin also plays a critical role in regulating the cell cycle at mitosis. To prevent apoptosis, survivin interacts with many regulatory factors [10]. Survivin is a small protein (displaying approximately 16.5-kDa band on SDS electrophoresis) and contains a single baculovirus IAP repeat but no RING finger motif, found in other IAP protein members. It is generally thought that transcriptional deregulation is a major mechanism involved in the aberrant expression of survivin in cancers. Transcription of the survivin gene proceeds from a single promoter but 5 splicing isoforms arise from the primary transcript [3]. The common survivin isoform contains 4 exons and isoform ΔEx3 (lacking exon 3) interacts with 4-exon survivin in the mitochondria where they inhibit mitochondrial-dependent apoptosis. Other survivin splice variants also colocalize with survivin in the mitochondria [3].

Despite the fact that almost 20 years elapsed from the discovery of survivin [11] and the mechanisms of its anti-apoptotic activities were widely studied [4], only little is known about its transcriptional regulation and maintenance of the high survivin levels is tumors. Here we briefly review the mechanisms of survivin transcriptional upregulation in the tumor tissue.

The survivin promoter harbors binding sites for several prooncogenic transcription factors (Sp1, STAT3, NF-kB, KLF5, E2F1, GATA1, DEC1 or TCF4) [6] that may be important in the elevation of survivin levels specifically in individual tumors. It is not known whether these sites could render sufficient survivin promoter activity required for the high survivin expression observed in all malignant cells. The most important transcription factors which maintain the survivin basal expression are Sp1 and Sp3. Several putative G/C-rich Sp-binding sites are present in the TATA-less survivin proximal promoter, two of which have been recognized as essential in regulating basal promoter activity [12]. Sp1-4 factors may play a role in activating some signaling pathways and receptors and contribute to oncogenic properties of cancer cells [13]. Attenuation of Wnt/β-catenin signaling resulted in inhibition of survivin expression in several cell types, suggesting a positive regulation of survivin expression by this pathway in the tested cell lines [14,15]. Herceptin inhibited survivin expression through the ErbB2-β-catenin/TCF4 pathway in breast cancer cells, completely dependent on TCF-4 binding sites in the promoter [16]. CBP (CREB-binding protein) seems to be a crucial cofactor for β- catenin/TCF-mediated upregulation of survivin, whereas CBP paralog p300 protein has an inhibitory effect [17].

Another upregulated signaling in cancer is the PI3K/Akt/mTOR pathway, which has been demonstrated to activate survivin expression in small cell lung cancer (SCLC) [18]. IGF-I (insulin-like growth factor-I) activated survivin expression in prostate cancer cells [19,20]. KLF5 transcription factor displayed positive effect on survivin expression in leukemia cells [21], and in ovarian CSC (cancer stem cells) [22]. In chondrocytes, DEC1-mediated apoptosis protection was achieved through the activation of survivin expression. DEC1 was demonstrated to activate survivin promoter, supported by two Sp1 sites [23]. Further, E2F1 protein positively correlated with survivin expression [24] and E2F1-3 transcription factors increased survivin expression in embryonic fibroblasts [25]. SOX2, a stem cell factor, directly upregulated survivin expression [26]. Hypoxia has also been reported to upregulate survivin and its promoter activity [27]. Pokemon is a pro-oncogenic transcription factor that was shown to bind the survivin promoter through the GT boxes and promoted survivin expression in MCF-7 breast cancer cells [28]. Interestingly, among various subtypes of non-small cell lung cancer (NSCLC), the survivin expression was higher in squamous cell carcinomas than in adenocarcinomas [29]. CK2 (casein kinase 2) promotes the survivin expression through the upregulation of β-catenin-dependent transcription [30,31]. Since the survivin promoter is also regulated negatively, mainly through the p53 protein, the positive transactivation effect of the HPV protein E6 on survivin is mediated via the p53 degradation [32]. Survivin expression was completely abolished by the knockdown of BRG1, a critical ATPase of the SWI/SNF-chromatin remodeling complex, indicating a direct regulation of survivin by an epigenetic mechanism [33]. Interestingly, Xu et al. [2] identified several polymorphisms in the survivin gene promoter, one of which is located at the repressor CDE/CHR element. This mutation was common among cancer cell lines, resulting in increased survivin transcription. Is has also been reported that the treatment of cancer cells with Hoechst33342, an AT-rich DNA-binding agent, upregulated survivin expression and promoter activity through the AT-rich DNA element in the survivin promoter [34].

Survivin expression, Hedgehog/GLI signaling pathway, and cancer stem cells

Recent studies have identified that the Hedgehog/GLI signaling cascade activates survivin transcription. We have identified many potential GLI-binding sites in the survivin promoter [35]. All sites contained 1-3 mismatched bases. However, it has been shown that these non-consensus GLI motifs are bound by GLI with similar affinity as consensus sites and lead to a strong transcriptional activation [36]. Across a wide panel of tumor cell lines, we found by using GANT61, a low molecular weight inhibitor of GLI factors, that survivin expression depends on GLI2 in at least one-half of tumor cell lines tested [35]. GLI2 associated with survivin promoter in chromatin immunoprecipitation assays and the expression of endogenous survivin could be easily evoked by ectopic GLI2 in human fibroblasts IMR90, where both survivin and GLI2 were normally silent. These findings have been supported by immunohistochemical staining, revealing correlation of GLI2 and survivin reactivity in tumor areas with a strong expression of both proteins [35]. Thus, survivin is a transcriptional target of GLI2 in a high proportion of various cancer cell lines including melanoma, NSCLC, SCLC, pancreas, and colorectal cancer and these findings may constitute a more general mechanism of survivin upregulation in tumors.

Many other target genes of Hedgehog/GLI pathway have been declared [37] but these might represent targets only in limited specific cell contexts. Survivin was described as a therapeutic target in Sonic Hedgehog-driven medulloblastoma [38], a tumor with invariant aberration of Hedgehog/GLI signaling, supporting our study. GANT61 was shown to induce caspase-independent apoptosis of SK-N-LO cells accompanied with a decrease of survivin [39]. The Hedgehog/GLI effector GLI1 is activated in malignant pleural mesothelioma and can be inhibited by SMO (Smoothened) inhibitors. SMO is a critical Hedgehog pathway membrane component. This SMO inhibition or GLI1 silencing resulted in reduced tumor growth and decreased survivin levels [40]. Overwhelming evidence points towards additional, noncanonical mechanisms of GLI1/2 factors activation, thus bypassing the necessity of upstream ligand signaling. Many cancer-deregulated pathways can directly activate GLI factors in tumor cells [41,42]. Therefore, GLI activity is not strictly dependent on the upstream SMO activation in tumor cells. Upregulated GLI proteins are associated with the invasive potential, epithelial-mesenchymal transition (EMT) and stemness of tumors [43,44]. Moreover, GLI2 is a direct activator of SOX2 transcription in telencephalic neuroepithelial cells and SOX2 expression decreases in the developing neuroepithelium of Gli2-deficient mice [4,45]. Additionally, it has been shown that SOX2 regulates self-renewal and tumorigenicity of human melanoma-initiating cells [46]. Of note, Notch1 cooperates with survivin to maintain stemness and stimulates proliferation of human keratinocytes [47]. Taken together, survivin, in a large part through the Hedgehog/GLI pathway, substantially contribute to the pro-oncogenic activities including stemness in human cancer.

It has long been known that a low or null survivin expression in normal tissues is maintained predominantly by the p53 protein [48-50]. This is mediated by the p53 target protein p21 [51]. If p21/ CDKN1A gene is deleted, p53-mediated repression is abolished [51]. Unlike Notch1, Notch2 inhibited survivin transcription [52]. Further, Egr-1 (early growth response 1 transcription factor) is also a negative regulator of survivin transcription. Its repression is direct, as Egr-1 binds to the promoter in vitro and in cells [53]. Similarly, KLF4 transcription factor is capable of binding the survivin promoter and repressing the expression of the survivin gene [54]. Caveolin-1 inhibits expression of survivin via a transcriptional mechanism involving the beta-catenin-TCF pathway, dependent on E-cadherin [55]. Tumor suppressor PTEN, a dual specificity phosphatase which represses the AKT/mTOR pathway, also silences the expression of survivin, independently of p53. The repression involves direct occupancy of the survivin promoter by factors FOXO1 and FOXO3a [56]. Peculiar regulation of survivin expression was observed in normal melanocytes which express survivin. While E2F transcription factors normally upregulate survivin transcription (above), E2F2 is a negative regulator of survivin expression in melanocytes. This repression requires active Rb protein. On the other hand, in the absence of activating stimuli, p53 and Rb proteins repress survivin transcription in normal melanocytes [7]. Yang et al. have demonstrated that TGF-β rapidly downregulates survivin expression in prostate epithelial cells. This repression involves Rb/E2F4 repressive complex and the CDE/CHR motif in the survivin promoter which acts as repressor element [57].

Because the inhibition of survivin expression reduces the tumor cell phenotype and because survivin may be a valid cancer therapeutic target, experiments were made to inhibit the survivin promoter by various agents. Hedamycin, a GC-rich DNA binding drug, downregulated survivin expression. Hedamycin-DNA interaction abrogated the binding of Sp1 to the survivin promoter and decreased its activity [58]. Further, selenium, a chemopreventive agent for many cancers, downregulated survivin expression similarly as hedamycin, i.e. via binding to the specific DNA site and preventing the association of Sp1 with DNA in the survivin promoter [59]. Anti-cancer antibiotics mithramycin A and doxorubicin also downregulated the endogenous survivin gene expression [60]. Since ras oncogenes augment the survivin levels, ras inhibitor farnesil thiosalicylic acid has been tested in glioblastoma multiforme U87 cells. It substantially decreased the survivin protein level and induced caspase-dependent cell death [61]. Notably, natural flavonoid compound wogonin induced apoptosis accompanied by a significant decrease of survivin and increase of Bax and p53. LY294002, a specific PI3K inhibitor, significantly accelerated wogonin-induced cell apoptosis [62]. Importantly, a specific survivin suppressant YM155 (sepantronium bromide), already tested in clinical trials, robustly inhibits survivin activity at low concentrations and is a promising drug for the targeted therapy of tumors. YM155 inhibits survivin activity via disruption of Sp1-DNA interaction in the survivin core promoter [63].

Tumor-specific survivin promoter can be used in gene therapy

As survivin expression is strictly confined to the tumor tissue, attempts have been made to utilize the survivin promoter in vectors carrying tumor-deleterious agents such as oncolytic viruses in both in vitro and in vivo studies. The survivin promoter specificity has been verified by expressing SEAP (secreted alkaline phosphatase), showing no expression in normal cells and tissues and high expression in tumors in cell culture and in mice [1]. The tumor specificity of survivin expression was also tested in lung cancer cells with the vector expressing a luciferase gene driven by the survivin promoter and evaluated in vitro and in vivo [64]. High specificity for tumor tissue was observed, indicating that survivin promoter is a cancer-specific promoter and may be useful in cancer gene therapy. Lu at al. [65] tested four promoters for tumor cell specificity in melanoma cell lines and found survivin promoter as the best among others. Conditionally replicating survivin promoter-responsive adenovirus (CRA) exhibited cancer-selective replication and induction of cell death. Survivin-CRA efficiently replicated and potently induced cell death in most types of cancer whereas the minimal viral replication in normal cells did not induce any detectable cytotoxicity in normal cells [66]. Similarly, recombinant adenoviral vector in which survivin promoter and a luciferase reporter gene were inserted into the E1-deleted region of the adenovirus vector exerted results positive only for the tumor tissue [67]. A specific adenoviral vector, in which E1A and E4 protein expression was dependent on survivin promoter activity, was replicated more than 100-fold better in bladder tumor cells than in normal cells [68]. It implies that the survivin promoter may be utilized for the construction of a replication-competent adenovirus to treat bladder cancers. The last three studies [66-68] were performed both in vitro and in vivo . Taken together, survivin promoter may prove to be a good candidate for transcriptional targeting in cancer gene therapy of a variety of tumors.

The data showing that survivin expression and Hedgehog signaling pathway are linked through the transcription factor GLI2, which activate endogenous survivin expression and can ectopically evoke survivin in normal, survivin non-expressing cells [35], could open new ways of tumor therapy. Chemical inhibitors of GLI2 and survivin are available and may act in synergy during treatment. In some tumor cells transcription factors other than GLI2 may keep the high survivin expression, dependent on the cell context. This might be highly individual for each tumor. Furthermore, survivin promoter has been shown to be highly specific for tumor cells and tissue in vitro and in vivo in several studies [1,64-68], having a therapeutic potential in gene therapy. So, we ought to continue exploring survivin protein as a target of treatment and survivin promoter as a carrier of deleterious agents to tumor cells.

This study was supported by funding from IGA, Ministry of Health of the Czech Republic (grant NT/14005-3 to JV) and from the Institutional research project PRVOUK-P25/LF1/2 from the Charles University, Prague, Czech Republic. The authors apologize to all the authors whose work could not be cited for reasons of an overwhelming number of survivin literature.