Introduction
Bladder cancer is the ninth most frequently diagnosed cancer and the 12th leading cause of deaths worldwide. It is the second most common urinary tract malignancy and the fifth most common cancer in men and the 19th in women [
1]. Of these tumors, 95 % are transitional cell carcinoma (TCC). Muscle invasive bladder cancer is related with high frequency of metastasis. Approximately 80 % of bladder cancers are nonmuscle invasive bladder cancer that rarely progress, and patients have good prognosis, but 30 % of those tumors progress into more aggressive and lethal forms. It is important to understand the molecular mechanism of bladder cancer metastasis to prevent cancer’s spread or to detect new therapeutic targets.
Epihtelial-mesenchymal transition (EMT) is a process by which epithelial cells lose their epithelial properties and obtain a mesenchymal phenotype. Tumor cells undergo epithelial to mesenchymal transition which transforms them from a quiescent cancer cells to a malignant phenotype. Loss of E-cadherin expression and induction of N-cadherin expression are a hallmark of the EMT process, which is needed for epithelial cells to adopt mesenchymal characteristic, a process also known as the cadherin switch [
2]. A reduction or loss in expression of E-cadherin has been associated with bladder tumorigenesis. Supression of E-cadherin expression by transcriptional factors, including Snail, Twist, and Zeb, is engaged in various malignancies.
We have previously demonstrated that EMT markers in melanoma cells are dependent on ILK function [
3]. ILK is a multifunctional intracellular effector of cell-matrix interactions and controls many cellular processes, including proliferation, survival, differentiation, migration, and invasion. ILK is a serine-threonine protein kinase which interacts directly with cytoplasmic domains of the β
1 or β
3 integrin subunits [
4]. ILK coordinates several signaling pathways. In particular, it phosphorylates and activates Akt at Ser 473 which controls the genes essential for survival [
4]. ILK can also directly phosphorylates GSK-3β at Ser 9, inactivate it, and lead to activation of some transcription factors [
5]. Overexpression of ILK leads to downregulation of E-cadherin and nuclear accumulation of β-catenin and NF-κB activating the expression of other mesenchymal genes [
3,
5‐
7]. The mechanism by which ILK induces the loss of E-cadherin and the progression of EMT is still unexplored, but current data suggest that ILK transcriptionally regulates Snail [
8] or through Poly (ADP-ribose) polymerase-1 (PARP-1), [
9] an unknown mechanism.
The mechanism associated with the role of ILK in tumor progression of bladder cancer is not well understood. Gao et al. [
10] presented that ILK is involved in bladder cancer cell proliferation, growth, and apoptosis, and Matsui et al. [
11] suggest that ILK expression is up regulated in invasive bladder cancer and plays a significant role in the EMT of bladder cancer by the control of E-cadherin and MMP-9 expression. They indicate that ILK regulates the EMT of bladder cancer and the mechanism depends on cell types. Although ILK is believed to be a key factor in EMT, its role in bladder cancer progression is not completely understood. The aim of the present study was to elucidate the mechanism of ILK-induced EMT and cadherin switch as a hallmark, in bladder cancer cells.
Discussion
The epithelial-mesenchymal transition gives mesenchymal properties on epithelial cells and is closely linked with development of aggressive traits by cancer cells. A pivotal alteration that occurs during EMT is the “cadherin switch”, in which the expression of E-cadherin is substituted by abnormal expression of N-cadherin. This downregulation of epithelial cadherin is linked with the release of β-catenin, which then travels to the nucleus and activates transcription of other genes responsive for EMT.
A large amount of research demonstrate interactions between cadherins and integrins, suggesting a connection between signaling pathways both integrin and cadherin. Alexander et al. [
17] presented that the induction of N-cadherin mRNA is dependent on β
1 integrin in PC-3 prostate adenocarcinoma cell line. In addition, in mouse mammary epithelial cells, Shintani et al. [
18] documented that upregulation of N-cadherin and EMT is initiated by collagen I receptors. Koenig et al. [
19] evidenced that collagen type I by interaction with β
1 integrins caused loss of E-cadherin-mediated cell-cell contacts and promoted proliferation of pancreatic carcinoma cells. Additionally, Kim et al. [
20] reported that deletion of α
3β
1 integrin in epithelial cells prevented EMT response to TGF-β. Our data demonstrate a new molecular mechanism in which ILK can regulate the “cadherin switch”. In our earlier work, we suggested that ILK may mediate cross-talk between adhesion molecules in melanoma cells and the control of cadherin expression by integrins might be a general feature common to all motile cells [
3].
It is known that ILK suppresses E-cadherin expression by controling the expression of its suppressor, Snail [
5]. Overexpression of ILK was associated with the expression of the E-cadherin repressor Snail and N-cadherin in pancreatic adenocarcinoma [
21]. Matsui et al. [
11] presented that ILK overexpression correlates with bladder invasiveness via the control of E-cadherin and matrix metalloprotease 9 (MMP-9). We also noticed that ILK can regulate expression of MMP-9 (Supplementary Fig.
1). Activation of ILK causes the regulation of numerous signaling pathways that in turn regulate EMT. We showed that knockdown of ILK diminished phosphorylation of downstream signaling target protein kinase Akt and glycogen synthase kinase -3β (GSK-3β), whereas the levels of Akt and GSK-3β protein remained almost unchanged. GSK activity is regulated through both serine and tyrosine phosphorylation [
22]. Akt phosphorylates GSK-3β at Ser 9, leading to inactivation of its kinase activity, and ILK can phosphorylate GSK-3β at Ser 9 in Akt-independent way [
23]. Luo et al. [
24] demonstrated that ILK knockdown inhibits EMT through inhibition of GSK-3β phosphorylation at Ser 9. On the other hand, actions which promote cell death, such as growth factor removal, stimulate kinase activity by increasign phosphorylation within the catalytic domain at Tyr 216. On top of it, Meares and Jope [
12] presented that phosphorylation of Tyr 216 of GSK-3β is responsible for the nuclear distribution of GSK-3β. They demonstrated that upon mutation of Tyr 216 to Phe, less GSK-3β is shifted to the nucleus. ILK knockdown in this study decreased the levels of phospho-Ser 9 GSK-3β form with a significant increase in the level of GSK-3β phospho-Tyr 216 form. We had similar results after silencing the ILK in melanoma cells [
3]. This was accompanied by significant decrease in expression of β-catenin, compared to the control. It is well known that Akt, GSK-3, and β-catenin are important mediators of ILK signaling. Yang et al. [
25] described the similar observation that ILK can regulate β-catenin accumulation in the nucleus and its activation in epithelial cells or intercellular communication via gap junctions. β-catenin plays also a critical structural role in cadherin-based adhesion. Although E-cadherin-bound β-catenin is relatively stable, the availability of β-catenin for binding cadherins is regulated by its phosphorylation [
13]. The cytosolic pool of β-catenin is also regulated by a phosphorylation-based mechanism. β-catenin is phosphorylated by GSK-3β at Ser 33, Ser 37, and Thr 47 within the N-terminal domain, which leads to its degradation by the ubiquitin/proteasome pathway. Our data show that siRNA-mediated depletion of ILK caused a phosphorylation on N-terminal residues and decrease in the total level of β-catenin in both bladder cell lines. These studies indicate ILK as a critical factor for nuclear translocation of β-catenin and activation of transcription factors, which upregulate the expression of oncogenic and mesenchymal genes in bladder cells. Furthermore, Fang et al. [
13] documented that β-catenin accumulation in the nucleus is not sufficient for β-catenin/TCF transcriptional activity. They showed that cells stably transfected with β-catenin S552A had a lower transcriptional activity that has been linked to tumor cell invasiveness. They proved that phosphorylation of β-catenin at C-terminal residues Ser 552 by Akt was necessary for the promotion of β-catenin transcriptional activity. Additionally, Miyabayashi and collegues [
26] suggested that phosphorylation of β-catenin may dictate to which particular co-activators bound to β-catenin and as a result, target genes can be activated. Our results also indicate the essential role of phosphorylation in the regulation of β-catenin function, based on the data that phosphorylation of β-catenin at Ser 552 is eliminated upon ILK silencing in both cell lines. β-catenin transcriptional activity has been associated with tumor progression, and nuclear translocation of β-catenin is very often used as an EMT marker. Our observation that ILK regulates expression of either EMT-inducing transcription factors Twsit-1, Zeb-1, and Snail indicates that ILK may function as upstream regulator of EMT associated signaling networks. Twist-1, Zeb-1, and Snail are zinc-finger transcriptional repressors that bind directly to the E-boxes of the promoter of the E-cadherin coding gene, and Twist-1 has been suggested as the major regulator of N-cadherin expression during gastrulation in
Drosophila and promoting the N-cadherin expression in gastric cancers [
27]. Our data indicated that regulation of cadherin switch by ILK pathway was mediated by the transcriptional repressors, as the loss of nuclear translocation of Snail, Zeb-1, and Twist-1 leads to upregulation of E-cadherin. E-cadherin re-expression also supports relocalization of β-catenin from the nucleus to the plasma membrane.
Several investigators reported new insights to control the production, stability, and intracellular localization of these transcriptional repressors [
9,
15,
28,
29].
McPhee et al. [
9] took notice that PARP-1 binds the SIRE sequence in Snail-1 promoter only in the presence but not the absence of ILK. They also showed that loss of ILK expression in prostate cancer cells had no effect on PARP-1 protein levels. The date presented in this study indicated that the amount of PARP-1 protein was reduced when ILK expression was inhibited. Snail activity can also be controlled by GSK-3. Zhou et al. [
28] reported that Snail-1 is phosphorylated by GSK-3 on two distinct motifs. Phosphorylation of two serines in the first motif leads to Snail ubiquitination, whereas phosphorylation of four serines on the second motif directs nuclear export. Mutation of all six serines increased the half-life of the Snail which resides exclusively in the nucleus to induce EMT. We did not observe detectable levels of Snail in nuclear fraction of HCV29 nonmalignant transitional epithelial cells, and we have shown inhibition of nuclear translocation of Snail in T24 bladder transitional cancer cells after silencing of ILK. Thus, Snail and GSK-3β together function as a molecular switch for many ILK signaling pathways that lead to EMT. We demonstrated that targeting ILK could control also the nuclear translocation of Zeb-1 and Twist in T24 bladder transitional cancer cells. Wu et al. [
29] suggested that Zeb-1 expression is necessary for transitional cancer cell invasion and distant metastasis in bladder cancer and that β-catenin induces Zeb-1 transcription. They also found that Zeb-1 could regulate expression of cytokeratins, vimentin, and MMP-2 but not N-cadherin expression. Although the mechanisms regulating the aberrant expression of N-cadherin in carcinoma progression remain unknown, the Twist-1 expression has been indicated to be necessary for N-cadherin expression during gastrulation in
Drosophila. Yang and colleagues suggested that Twist expression was sufficient to induce in vitro EMT [
27], and Alexander and colleagues demonstrated that integrin-mediated adhesion is involved in the Twist-1 nuclear translocation and is necessary for N-cadherin expression in PC-3 prostate carcinoma cells [
17]. Recently, Yang and coauthors discovered that knocking down ILK or inhibiting FAK, MAPK/ERK, or PI3K/Akt signaling suppressed Twist-induced EMT [
30]. Notably, they found that integrin β
1 acts as a core regulator in this network. Our date confirmed that ILK is implicated in nuclear translocation of Twist, because Twist was absent in the nucleus after ILK silencing in comparison to control cells in T24 cells. Although the total level of Twist-1 did not change significantly, Twist protein was present only in cytoplasm fraction. Our results showed that in T24 cells, the alteration of protein level was more significant than that of the mRNA of N-cadherin, suggesting posttranscriptional regulation of N-cadherin by Twist, as it has been recently observed in gastric cancer [
27]. Twist protein is able to form homo and heterodimers, and posttranslational modification, such as phosphorylation, can alter its interaction with other proteins and the binding to DNA. In our previous work, after ILK silencing in melanoma cells, we observed a decrease in expression of N-cadherin but only at the protein level; the re-expression of E-cadherin on mRNA or protein level was not observed [
3]. The specific mechanism through ILK modulates N-cadherin expression is not clear. Presented results show that ILK pathway regulates the cadherin switch of bladder cancer through multiple mechanisms, including transcriptional and posttranslational regulation.
It is difficult to indicate a common pathway controlling EMT under ILK regulation, assuming that various pathways may be involved and may vary in different cell types.