miR-145 induces caspase-dependent and -independent cell death in urothelial cancer cell lines with targeting of an expression signature present in Ta bladder tumors

Abstract

Downregulation of miR-145 in a variety of cancers suggests a possible tumor suppressor function for this microRNA. Here, we show that miR-145 expression is reduced in bladder cancer and urothelial carcinoma in situ, compared with normal urothelium, using transcription profiling and in situ hybridization. Ectopic expression of miR-145 induced extensive apoptosis in urothelial carcinoma cell lines (T24 and SW780) as characterized by caspase activation, nuclear condensation and fragmentation, cellular shrinkage, and detachment. However, cell death also proceeded upon caspase inhibition by the pharmacological inhibitor zVAD-fmk and ectopic expression of anti-apoptotic Bcl-2, indicating the activation of an alternative caspase-independent death pathway. Microarray analysis of transcript levels in T24 cells, before the onset of cell death, showed destabilization of mRNAs enriched for miR-145 7mer target sites. Among these, direct targeting of CBFB, PPP3CA, and CLINT1 was confirmed by a luciferase reporter assay. Notably, a 22-gene signature targeted on enforced miR-145 expression in T24 cells was significantly (P<0.00003) upregulated in 55 Ta bladder tumors with concomitant reduction of miR-145. Our data indicate that reduction in miR-145 expression may provide bladder cancer cells with a selective advantage by inhibition of cell death otherwise triggered in malignant cells.

Introduction

MicroRNAs (miRNAs) are small non-coding RNA molecules that regulate gene expression post-transcriptionally in multi-cellular organisms by interaction with partially complementary target sites in mRNA molecules. In accordance, miRNAs have been shown to influence gene regulatory processes in, for example, development, differentiation, and disease, such as cancer (Calin and Croce, 2006; Esquela-Kerscher and Slack, 2006), in which specific miRNA signatures have been associated with different clinical outcomes (Nakajima et al., 2006; Rajewsky, 2006; Yanaihara et al., 2006; Sempere et al., 2007; Tavazoie et al., 2008; Dyrskjot et al., 2009). Approximately 50% of the known miRNAs are located within regions of genomic instability that are amplified or deleted in cancer (Calin et al., 2004), which may in part explain the aberrant expression observed. MIR145 is located at chromosome 5q32 in a putative bicistronic cluster with MIR143 (Cordes et al., 2009), a region commonly lost in myelodysplastic syndromes (Horrigan et al., 1996).

miRNAs are predominantly transcribed by polymerase II and processed by Drosha into 70–100 nt precursor molecules (Lee et al., 2003). Once exported out of the nucleus by Exportin-5, pre-miRNAs are processed into 22 nt duplexes and on separation, single-stranded mature miRNAs interact with target mRNAs. According to the current dogma, miRNAs interact predominantly with sequences in the 3′ untranslated region (UTR) of mRNAs that are complementary to nt 2–8 of the miRNA (termed the ‘seed’) resulting in mRNA destabilization and/or translational repression depending on the degree of complementarity, and under the influence of RNA-binding proteins, such as Dnd1 (Kedde et al., 2007; Baek et al., 2008). However, miRNAs may also enhance translation on binding in the 5′ UTR and oscillate between mediating translational repression or activation (Vasudevan et al., 2007; Orom et al., 2008).

Reduced expression of miR-145 has been observed in a variety of cancers, such as colon, lung, ovarian, prostate, and breast cancer (Michael et al., 2003; Iorio et al., 2005; Yanaihara et al., 2006; Porkka et al., 2007; Yang et al., 2008). The downregulation is evident already in precancerous stages of breast and colon cancer (Michael et al., 2003; Sempere et al., 2007), suggesting loss of miR-145 as a common early neoplastic event. In situ hybridization (ISH) analysis has identified miR-145 expression in myoepithelial cells of the mammary lobules and ducts as well as in smooth muscle cells (Sempere et al., 2007; Cordes et al., 2009). Furthermore, studies have shown that miR-145 reduces cell viability in colon and cervical cancer cell lines after prolonged exposure (Shi et al., 2007; Schepeler et al., 2008; Wang et al., 2008). Finally, the expression of miR-145 has been shown to be induced by p53 (Sachdeva et al., 2009) and regulated by an enhancer element activated by SRF and Nkx-2 transcription factors (Cordes et al., 2009).

We recently identified miR-145 as the most significantly downregulated miRNA (P=2.1E−08) in an miRNA profiling study on bladder cancer comprising 106 clinical samples of bladder cancer and 11 normal bladder biopsies (Dyrskjot et al., 2009). Bladder cancer is the fourth most common cancer among men and can be pathologically classified into discrete stages; Ta are benign papillary tumors, T1 are lamina propria-invasive tumors, carcinoma in situ (CIS) are small in situ lesions of high-grade atypia, and T2–T4 are muscle-invasive tumors. In this study, we used two different approaches to examine the possible function of miR-145 in bladder cancer cells. First, we investigated the localization of miR-145 in clinical samples and the phenotypic consequence of miR-145 expression in cell culture with identification of putative targets involved in the cellular signaling. Second, a bioinformatic analysis was conducted to clarify if the identified target genes could possibly be under the regulatory influence of the miR-145 level in clinical specimens and thereby increase the understanding for its reduction in bladder cancer.

Results

Localization of miR-145 in clinical samples by ISH analysis

The downregulation of miR-145 (Figure 1a) in bladder cancer specimens and urothelial cell lines, as compared with normal biopsies (Figures 1b and c) (Dyrskjot et al., 2009), prompted us to investigate whether the differences in miRNA levels measured may reflect tissue heterogeneity and/or an altered ratio of different cell types to the tumor mass, rather than tumor cell-specific expression changes. We performed ISH analysis to identify which cell type(s) expressed miR-145 (Figure 2). Detection of miRNAs by ISH is challenging because of their small size, we, therefore, applied an improved method using locked nucleic acid (LNA)-modified probes to ensure hybridization to single-stranded RNA molecules (Nuovo, 2008). In normal tissue biopsies from bladder, the miR-145 probe gave a strong signal in the urothelium (Figure 2a, bottom panel). The uppermost umbrella cell layer displayed less signal than the other layers of urothelium. Infiltrating lymphocytes in the stroma, endothelial cells of blood vessels, and the muscularis mucosa also displayed miR-145 expression. Papillary Ta tumors showed homogenous expression in carcinoma cells with reduced staining intensity compared with normal urothelium, and muscle-invasive tumors showed heterogeneous expression among the carcinoma cells (Figure 2b), usually with large carcinoma cell areas absent of signal (arrow) and occasional small ‘islets’ of intense signal (asterisks). In CIS lesions, the miR-145 signal was predominantly absent (Figure 2b). The localization examined on a tissue microarray containing core biopsies from 182 Ta, 101 T1, and 34 T2–T4 tumors gave similar results. Here, reduced expression of miR-145 in T1 high-grade tumors correlated with disease progression with borderline significance (P=0.057, Figure 2c).

Figure 1
figure1

miR-145 expression in clinical samples of normal bladder biopsies, bladder cancer, and in bladder cancer cell lines. (a) Secondary structure of premiR-145, the sequence of the mature miR-145 is shown in lower case, and the 7mer seed sequence is boxed. (b) Different stages of bladder cancer. (c) Expression pattern of miR-145 in 11 normal bladder biopsies (N), 27 superficial Ta tumors (Ta), 40 T1 tumors (T1), and 27 invasive T2–T4 tumors (T2–T4), and bladder cell lines: non-malignant epithelial cell lines HU609 and HCV29, transitional cell carcinoma cell lines T24, SW780, HT1376, J82, and transitional cell papilloma cell line RT4. Samples were analyzed by LNA-based oligonucleotide microarray (y axis, log2 scale).

Figure 2
figure2

miR-145 localization in clinical samples by ISH. (a) A representative normal bladder biopsy, (scalebar-100 μm). Upper panel: HE staining, middle panel: negative control ISH using a mismatch LNA-probe (miR-145 mm), no staining observed, bottom panel: ISH using a perfect match LNA-probe (miR-145). (b) A representative CIS lesion and T2 tumor (scale bar, 200 μm). (Arrow, area absent of signal; asterisks, small ‘islets’ of carcinoma cells of intense signal). (Uro, urothelium; Umbr, umbrella cells; Endo, endothelial cells; M, muscle tissue; Carc, carcinoma cells). (c) Kaplan–Meier plot of the probability of progression-free survival among 94 patients with T1 high-grade tumors based on the miR-145 ISH signal in carcinoma cells in core biopsies of tumors on a tissue microarray.

Time- and dose-dependent cell death in bladder cancer cell lines on exogenous miR-145 expression

To analyze the possible tumor suppressor function of miR-145, we transfected miR-145 or scrambled (scr) precursor molecules into the urothelial carcinoma cell lines T24 and SW780, and as a control into HU609 cells. HU609 is a non-malignant urothelial cell line (Litynska et al., 2000) that does not induce tumors in mice (data not shown). Efficient transfection was confirmed (Supplementary Figure S1) and exogenous mature miR-145 was detected already at 12 h post-transfection (Figure 3a). Exogenous expression of miR-145 has been shown to reduce the growth of HeLa cells by 30% and colon cancer cells 4 days post-transfection (Shi et al., 2007; Wang et al., 2008). In T24 and SW780, transfection of 10 or 50 nM pre-miR-145 exerted an antiproliferative effect that was time dependent with onset already at 24 h post-transfection, whereas immortalized HU609 cells were only slightly growth inhibited after 72 h (Figure 3b). Lower amounts of miR-145 (0.01–0.1 nM) did not significantly affect cell viability, whereas 1 nM moderately reduced it (Figure 3c). To test whether miR-145-induced reduction in cell density was associated with cell death, we analyzed whether miR-145 expression resulted in plasma membrane permeabilization. A dose-dependent release of the cytoplasmic enzyme, LDH, into the media was observed for the carcinoma cell lines using 10–50 nM pre-miR-145, but not for HU609 cells (Figure 3d). The cell death observed in T24 cells was confirmed by an increase in the number of cells positively stained with the cell-non-permeable dye SYTOX-Green (Supplementary Figure S1c).

Figure 3
figure3

The effect of exogenous miR-145 introduction on cell viability and cell death. (a) HU609, T24, and SW780 cells were transfected with 10 nM scr pre-miRNA or pre-miR-145. The level of mature miR-145 was examined 12, 24, and 48 h by RT–qPCR (note: excessive cell death was, however, observed at 48 h in SW780 cells). (b) HU609, T24, and SW780 cells were transfected with the indicated pre-miRNA and the time-dependent viability was determined by the MTT reduction assay and expressed as the viability compared with untransfected cells. (c) Dose-dependent viability 96 h post-transfection was examined using the indicated concentrations of pre-miRNAs. (d) Dose-dependent cell death was determined by the LDH release assay, and expressed as percentage of released LDH out of total cellular LDH. Columns (a), averages of a triplicate experiment, points (b) and columns (c, d), averages of triplicate experiments; bars, s.d. (ad). Data are representative of a minimum of three triplicate experiments.

Involvement of caspases during miR-145-induced cell death and the effect of exogenous miR-145 knockdown

Different programed cell death modes can be defined on the basis of cellular morphology (Kroemer et al., 2008). Necrosis-like programed cell death is characterized by swelling of the cell, early disruption of the plasma membrane, and lack of chromatin condensation, whereas hallmarks of apoptosis and apoptosis-like programed cell death are shrinkage of the cell, blebbing of the plasma membrane, and compact or loose chromatin condensation before the rupture of the plasma membrane, respectively. Autophagic cell death is characterized by massive accumulation of autophagosomes (Levine and Yuan, 2005). T24 and SW780 cells transfected with pre-miR-145 displayed cell rounding, shrinkage, plasma membrane blebbing, and subsequent detachment of the cells (Figure 4). The kinetics of the death process was cell type dependent; SW780 cells being in late stages of cell death by 48 h, whereas T24 cells displayed a flattened/senescent morphology (Figure 4a and data not shown). No phenotype change was observed for HU609 cells (data not shown). The morphological features of an apoptotic cell death were preceded by activation of effector caspases 3 and 7 as measured by kinetic cleavage of the substrate DEVD-AFC (Figure 4b, similar results were observed for SW780 cells), indicating that miR-145 induces cell death through caspase activation.

Figure 4
figure4

Cell morphology and caspase activation upon miR-145 expression. (a, c, d) T24 and SW780 cell lines were transfected with pre-miR-145 at the indicated concentration ±50 nM LNA antagonist, or treated with 50 μM etoposide (Eto) when indicated. (a) Cell morphology was examined by phase contrast microscopy 48 h post-transfection, (scale bar, 100 μm). (b) The effector caspase activity (DEVDase) in T24 cell lysates 24, 48, and 72 h post-transfection was examined by fluorometric kinetic analysis and normalized toward total cellular LDH. (c) The level of miR-145 was examined 48 h post-transfection by RT–qPCR. (d) Viability was examined in T24 and SW780 cells 48 h post-transfection. Columns (b d), average of triplicate experiments; bars, s.d. Data are representative for a minimum of three triplicate experiments. **P-value <0.01, -P-value >0.05.

We next performed a loss-of-function experiment by transfecting an LNA knockdown molecule against miR-145. This reduced the level of exogenous mature miR-145 (Figure 4c), restored cell morphology (Figure 4a), inhibited caspase activation (Figure 4b), and recovered cell viability (Figure 4d). These results strongly indicate that effector caspases participate in the execution of cell death induced by miR-145, and, furthermore, indicate activation of a classical apoptotic death mode.

Activation of an alternative cell death pathway upon blockage of the caspases

Rather surprisingly, we found that miR-145-induced cell death proceeded when SW780 cells were incubated with the broad caspase inhibitor zVAD-fmk, as determined by the morphological changes observed (Figure 5a, rounding up and loss of membrane integrity, asterisks) and the reduced viability 48 h post-transfection (Figure 5b). This indicates that caspase activation is not essential for miR-145-induced cell death in SW780 cells. Next, DNA staining with cell-permeable Hoechst-33342 and non-permeable SYTOX-Green that enables detection of nuclear changes and loss of plasma membrane integrity, respectively, was performed. The presence of compact condensed and fragmented Hoechst-stained nuclei that were SYTOX-Green negative indicated classical apoptotic nuclear changes occurring before plasma membrane disruption (Figure 5c, arrows and close up). In the presence of zVAD-fmk, loose nuclear condensation was observed without fragmentation (Figure 5c), and a higher fraction of these cells also displayed lost plasma membrane integrity (Figure 5c, graph) suggesting a shift from the classical apoptosis mode toward an alternative death pathway. Similar results were obtained for T24 cells (data not shown).

Figure 5
figure5

The function of caspase activation and Bcl-2 in miR-145 cell death signaling. SW780 cells were transfected with 50 nM scr pre-miRNA or 50 nM (a), 10 nM (b, left), or 10/50 nM pre-miR-145 (b, right), and, when indicated, subsequently treated with 50 μM zVAD-fmk (broad caspase inhibitor). (a) Morphology was examined by phase contrast microscopy 48 h post-transfection, (scale bar, 100 μm). (Asterisks, cells with lost membrane integrity). (b) Effector caspase activity in SW780 cell lysates at 48 and 72 h post-transfection and viability 48 h post-transfection. (c) Nuclear morphology 48 h post-transfection (50 nM). DNA was stained using cell-permeable Hoechst-33342 and non-permeable SYTOX-Green. Arrows, cells with condensed nuclei and intact plasma membrane (scale bar, 200 μm). High-resolution pictures of nuclei (scale bar, 20 μm). The percentage of SYTOX-Green positive condensed nuclei±zVAD-fmk is depicted in the graph. (d) T24 cells transfected with empty vector (pCEP4) or Bcl-2 (pCEP4-Bcl-2) were analyzed for expression of Bcl-2 by western blotting and the sensitivity was examined after transfection of 10 nM miR-145 (72 h post-transfection) or 10 μM etoposide (Eto) and 10 nM vincristine (Vin) (48 h) by MTT assay. Columns (bd), averages of triplicate experiments; bars, s.d. Data are representative of three experiments (b, d) or average of three experiments (c).

During classical apoptosis, pro-apoptotic proteins such as cytochrome c are released from the mitochondria; a process that triggers downstream caspase activation. This translocation is effectively inhibited by anti-apoptotic proteins of the Bcl-2 family (Ow et al., 2008). Stable expression of exogenous Bcl-2 in T24 cells was insufficient, however, to block the miR-145-induced cytotoxicity, whereas it conferred significant protection against etoposide and vincristine, two commonly used chemotherapeutics (Figure 5d). This indicates that miR-145-induced death signaling is independent of the Bcl-2 status and of signaling through the mitochondrial pathway.

Downregulation of mRNAs with 7mer target site(s) before the onset of miR-145-induced cell death

To identify putative targets for miR-145, we transfected T24 cells with 10 nM scr or miR-145 precursor and analyzed mRNA levels 46 h post-transfection using gene expression microarrays. The majority of mRNAs (85%) and miRNAs other than miR-145 (>95%) remained unchanged (Figure 6a and data not shown). The reported miR-145 targets, insulin receptor substrate-1, and MYC that have been speculated to influence the growth inhibitory effect of miR-145 (Shi et al., 2007; Sachdeva et al., 2009) were not expressed or in fact induced rather than repressed, respectively, and, therefore, unlikely to account for the phenotype observed in our model system. As individual miRNAs are estimated to target hundreds of mRNAs, we hypothesized that the phenotypic effects of miR-145 may be a consequence of the combined alteration of numerous targets instead of a single target gene and that it may result from both primary, secondary, and subsequent regulatory effects. Hence, we used a bioinformatic approach for our further analysis.

Figure 6
figure6

Gene expression changes induced by miR-145 and target gene identification. (a) RNA from scr- or pre-miR-145 transfected T24 cells was harvested at 46 h and subjected to exon array analysis. The table depicts the number of mRNAs either downregulated (‘−’), unchanged (‘=’), or upregulated (‘+’) in two independent samples (experiment 1 and 2) upon miR-145 transfection, as compared with the scr control. The average number of 7mer target sites (AACUGGA) for miR-145 in the mRNAs within each group is denoted in parenthesis. (Note the higher level of target sites in the downregulated genes, boxed). (b) Expression fold change of predicted miR-145 targets (TargetScan or PicTar), and of mRNAs±at least one 7mer site. (Note the higher level of downregulation among the predicted target genes). (c) Gene list of downregulated mRNAs identified on miR-145 overexpression and their expression in Ta bladder tumors compared with normal urothelium. The criteria for the identification of the 22-target gene signature: (1) only mRNA transcripts expressed above background level (above average signal intensity on the array) in untransfected cells were considered (10 786/22 433 entries). (2) Only genes present in the databases of the prediction algorithms Targetscan 4.2 and PicTar were considered (6249/10 786 entries). (3) Transcripts were classified as upregulated, non-differentially regulated or downregulated (up: >twofold and down : <twofold). Transcripts that were downregulated in at least one sample based on this criteria were selected (624/6249 entries). (4) The categories of mRNA transcripts were compared with the lists of predicted targets by Targetscan and PicTar. The categorized downregulated transcripts that were also predicted targets were identified (22/624 entries). The presence of 7mer target sites and the fold change in expression for the two experiments are listed. The right columns show the expression of these mRNAs, examined by U133A microarray analysis, in 55 Ta tumors (grade 1–3) compared with nine normal bladder biopsies (P-value, t-test, and fold change in expression) (note that as expected, there is an overall inverse relation between alterations in vitro and in vivo). (d) RT–qPCR validation of miR-145 predicted targets CLINT1, PPP3CA, and CBFB in T24 cells using GAPDH for normalization. Columns, averages of duplicate samples analyzed in triplicates; bars, s.d. (e) Luciferase-based target validation of CLINT1, PPP3CA, and CBFB. Luciferase constructs were fused to 400–800 bp of the 3′UTR containing the 7mer target site of CLINT1, PPP3CA, or CBFB. In parallel, constructs with a mutated target site were analyzed. Columns, average of triplicate experiments; bars, s.d.

mRNAs that were either unaffected or upregulated on miR-145 expression contained a 7mer target site at a frequency of 25–30% (Figure 6a), in agreement with the background frequency reported by others (Selbach et al., 2008). In contrast, a three- to fourfold enrichment of the 7mer was found among downregulated mRNAs as almost all of these harbored a 7mer site (P<1E-06, Figures 6a and b) and, interestingly, was not exclusively found in the 3′UTR, but also in protein coding regions (data not shown). We next asked if the 188 downregulated mRNAs (Supplementary Figure S2) in particular represented known biological pathways. Ingenuity Pathway analysis software designated significant enrichment for mRNAs involved in gene networks annotated with ‘cancer’, ‘cell cycle’, and ‘cell death’.

Predicted miR-145 targets (according to TargetScan and PicTar databases) and mRNAs harboring a 7mer site had a significantly higher propensity of being downregulated (Figure 6b). Bioinformatic analysis identified nine mRNAs that were (1) expressed above a background level in untreated cells, (2) downregulated at 46 h, (3) containing a 7mer target site, and (4) predicted targets (TargetScan/PicTar) (Figure 6c, gray shadow and S3). Consistent with the microarray data, the depletion of three of these, Clathrin Interactor 1 (CLINT1), core-binding factor β subunit (CBFB), and protein phosphatase 3 catalytic subunit α isoform (PPP3CA) was also observed by RT–qPCR (Figure 6d). Furthermore, reduced expression of a luciferase reporter was detected in cells co-transfected with pre-miR-145 when the luciferase construct carried 400–800 bp of the 3′UTR of CLINT1, PPP3CA, and CBFB containing the target site (Figure 6e). This effect was not observed using an scr pre-miR or on mutation of the target site. Finally, a 22-gene signature of miR-145 targets was generated, when loosening the selection criteria described above to contain mRNAs downregulated >twofold in at least one sample (Figure 6c, all 22 genes listed).

Investigation of the expression of miR-145-regulated mRNAs in Ta tumors

We next speculated, if the 22-target gene signature identified in our cell culture model was also present in clinical bladder cancer samples. We, therefore, validated the expression of the 22-gene signature in 55 papillary Ta tumors (18 grade I, 12 grade II, 25 grade III) and compared it with nine normal bladder biopsies by gene expression microarray analysis (Figure 6c). CBFB, PPP3CA, and 11 additional mRNAs of the signature were all significantly upregulated consistent with the low expression of miR-145 in these tumors (Figure 1). Using a cutoff of P<0.01 and fold change >2 for individual mRNAs, the likelihood of extracting this expression pattern by chance is extremely low (P<0.00003, binomial statistical analysis). A direct negative correlation between the expression of miR-145 and the three-target genes CBFB, PPP3CA, and CLINT1 was observed in eight clinical samples (Supplementary Figure S4). These data suggests that experimentally identified targets for miR-145 are also upregulated in Ta tumors consistent with a low miR-145 expression.

Discussion

In this study, we examined the function of miR-145, the most significantly downregulated miRNA in bladder cancer evident even in low stage disease. We showed that reduced miR-145 expression approached significance as a predictor of disease progression supporting the function as a tumor suppressor. In addition, we showed that miR-145 strongly induces cell death in two malignant urothelial cell lines, using more than one cell death pathway, but not in an immortalized non-tumorigenic cell line. Furthermore, we identified a set of target genes harboring the 7-mer target site and documented that this subset of transcripts is upregulated in clinical specimens with downregulated miR-145 hence establishing a correlation between the expression levels of miR-145 and its targets in a cell line model and human malignant urothelium.

The downregulation of miR-145 in bladder cancer biopsies led us to speculate, if this was caused by a mere variation in cellular composition. However, ISH analysis showed staining of miR-145 in the normal urothelium, and weak or heterogeneous expression in tumor samples, suggesting actual differences in miR-145 expression between normal and transformed urothelial cells. Consistent with this, a >100-fold decrease in miR-145 expression was detected by RT–qPCR in Ta tumors compared with normal bladder biopsies (data not shown). The mechanism of miR-145 downregulation in cancer is unknown, although defective processing of miR-145 has been suggested, as equal levels of the hairpin precursor, but not mature miR-145, has been observed in colon cancer (Michael et al., 2003). p53 response elements identified in the putative promoter region of mir145 suggest that upon p53 inactivation, the transcription of miR-145 may be reduced (Sachdeva et al., 2009). However, p53 is seldomly inactivated in Ta tumors (Mitra et al., 2007) and thus cannot account for the early loss of miR-145. We instead investigated if DNA hypermethylation could have a function in the reduction of miR-145 in Ta tumors. A significant overall increase in DNA methylation in a CpG-rich region 250 bp upstream of MIR145 in five Ta tumors (grade 1 and 2) compared with four normal bladder biopsies (P<0.0005, χ2-test, data not shown). Although few samples were analyzed, this suggests that DNA hypermethylation could contribute to the reduced expression, although other regulatory mechanisms are most likely also involved.

Earlier studies have shown that miR-145 reduces the viability of colon and cervical cancer cells 96 h post-transfection (Shi et al., 2007; Schepeler et al., 2008; Wang et al., 2008) and regulate a quiescent versus proliferative state of smooth muscle cells (Cordes et al., 2009). In this study, we identified miR-145 as an inducer of time- and dose-dependent massive cell death with classical apoptotic features, such as caspase activation, fragmentation and compact condensation of the chromatin, and plasma membrane blebbing. However, cell death was not inhibited by Bcl-2 or zVAD-fmk, indicating that an alternative cell death independent of the mitochondrial apoptosis pathway, and of the caspases, was also activated. Whether these pathways are activated in parallel or function in redundancy is unknown. The existence of an interplay between apoptotic and non-apoptotic cell death in cancer cells is well known (Kim et al., 2006). For example, TRAIL induces apoptosis in lung carcinoma A549 cells at normoxic levels and non-apoptotic death at hypoxia at which Bcl-2, Bcl-XL, and Inhibitor of apoptosis proteins are upregulated (Kim et al., 2004). In addition, cells are sensitized to TNF-α-induced cell death in the presence of zVAD-fmk that arrest apoptosis and promote other forms of cell death, such as autophagic cell death (Yu et al., 2004). As genetic alterations in cell death pathways enable cancer cells to evade cell death eventually leading to tumor progression (Hanahan and Weinberg, 2000; Kim et al., 2006), effective induction of multiple types of cell death may increase the therapeutic efficacy of anticancer agents. Thus, from a clinical therapeutic perspective, it is interesting that miR-145 induces multiple death pathways.

Although miRNAs target hundreds of transcripts, potentially leading to a cascade of secondary expression changes, our bioinformatic analysis identified the majority of mRNAs (85%) and other miRNAs (95%) as unchanged 46 h post-transfection. Our analysis also identified miR-145 as a predominantly negative regulator of gene expression rather than an inducer, as was unexpectedly reported for miR-10a (Orom et al., 2008). Among the downregulated mRNAs, Ingenuity Pathway Analysis identified significant enrichment of genes associated with ‘cell cycle’, ‘cell death’, and ‘cancer’ signaling, for example CBFB, PPP3CA, Ras p21 protein activator 1 (RASA1), homeodomain-interacting protein kinase 2 (HIPK2), X-linked Inhibitor of Apoptosis (XIAP/BIRC4), and the Proliferation-related Ki67-antigen (MKI67). Among these, CBFB is a subunit of the heterodimeric transcription factor CBF. A chromosomal abnormality, inv(16)(p13q22), found in 10% of acute myeloid leukemia cases results in the fusion of CBFB with MYH11, which generates the fusion oncoprotein CBFB-MYH11 (reviewed in Hart and Foroni, 2002). In Caenorhabditis elegans, overexpression of the CBFB homolog BRO-1 leads to massive hyperplasia (Kagoshima et al., 2007). These findings show that CBFB has a significant function in oncogenesis. We are currently investigating if the depletion of this and other targets may mimic the phenotype of miR-145 expression. Notably, we examined how these mRNAs were regulated in Ta papillary tumors. Consistent with our functional analysis, reduced miR-145 expression in Ta tumors correlated with a highly significant upregulation of the 22-target gene signature.

The miRNA array profiling and ISH analysis showed that miR-145 expression was decreased in low stage tumors and CIS lesions, whereas more heterogeneously expressed in T1 and T2–T4 tumors. An explanation for this shift could be that reduction of miR-145 is necessary to achieve a survival advantage during early oncogenesis, but not for later steps of carcinogenesis. miR-145 seems to be a tumor suppressor based on the induction of cell death and reduced expression on disease progression; therefore, inactivation of this tumor suppressor may be crucial for the initial triggering of an oncogenic pathway. For late tumor stages (T2–T4), additional mutations are acquired that affect the cellular homeostasis, which could explain why the loss of miR-145 not serves a selective advantage for tumor growth in these tumors. For example downstream effector molecules in miR-145-regulated pathways could be affected in these tumors relieving the pressure for a miR-145 loss.

One of the major challenges in the current treatment of bladder cancer is to identify and prevent recurrence and disease progression of patients with Ta tumors. Our data indicate that miR-145 is a potential candidate for future non-coding RNA medicine against this stage of bladder cancer with loss of miR-145, for example as a neoadjuvant therapy in combination with transurethral resection using intravesical installation for the aim of eliminating any remaining tumor cells or foci of CIS.

Materials and methods

Patient material

Biological material from bladder tumors was obtained directly from surgery and processed as described earlier (Dyrskjot et al., 2005). Informed written consent was obtained from all patients, and research protocols were approved by the scientific ethical committee of Aarhus County.

ISH detection of miR-145

Five μm sections FFPE-serial tissue sections were processed for ISH analysis essentially as described (Nuovo, 2008; Schepeler et al., 2008). An LNA-modified probe (Exiqon, Vedbaek, Denmark) complementary to miR-145 (probe: IndexTermagggattcctgggaaaactggac, ‘miR-145’) or with two mismatches to the seed sequence (probe: IndexTermagggattcctgggaaaaGtCgac, ‘miR-145 mm’) was used (For details see Supplementary information). In total, four normal biopsies, four Ta grade 1, −2, and 3, respectively, four T2–T4, and three CIS lesions as well as a tissue microarray containing 0.6 mm core biopsies from 189 Ta, 101 T1, and 34 T2–T4 tumors were analyzed by ISH. The signal intensity on the tissue microarray was scored blinded to clinical outcome by two independent observers (α statistics=0.71).

Cell culture and transfections

Human urinary bladder transitional cell carcinoma (T24, SW780, HT1376, J82, RT4) and immortalized human bladder epithelium (HU609, HCV29) cells were propagated in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10% heat-inactivated fetal calf serum and antibiotics at 37 °C in a humidified air atmosphere at 5% CO2. T24 pCEP4 and –pCEP4-Bcl-2 are stable cell lines transfected with an empty pCEP4-hygro vector (Invitrogen) or pCEP4-hygro encoding human Bcl-2 and propagated under selection (200 μg/ml hygromycin). HEK293 cells were propagated in RPMI-1640 containing 10% fetal calf serum and antibiotics. miRNAs (Ambion, Austin, TX, USA) and LNA knockdown molecules (Exiqon) were reverse transfected using Lipofectamine 2000 (Invitrogen) according to manufacturer's guidelines. The broad caspase inhibitor zVAD-fmk (Bachem, Bubendorf, Switzerland) was added 8 h after transfection. Transfection efficiency was examined on transfection of a Cy3-conjugated scr miR (Ambion) by fluorescence microscopy and flow cytometry.

miRNA target validation

A partial 3′UTR (dUTR) sequence of 400–800 bp from PPP3CA, CLINT1, and CBFB containing the potential miR-145 target was inserted to the XhoI/NotI site of the psiCHECK2 vector (Promega, Madison, WI, USA). In addition, two, one, and one putative, target sites were mutated for PPP3CA, CLINT1, and CBFB, respectively (see Supplemental information for primer sequences). HEK293 cells were co-transfected with 0.08 μg psiCHECK-UTR DNA and 50 nM pre-miRNA using Lipofectamine 2000. The luciferase activity was measured on a FLUOstar luminometer (BMG labtech, Offenburg, Germany) and normalized to the relative rluc/fluc value of the corresponding mutated UTR using Dual-luciferase reporter assay system (Promega).

RNA extraction and microarray analysis

RNA was extracted using a standard Trizol RNA extraction method (Invitrogen) and quality controlled using 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Microarrays for miRNA expression analysis were produced using an LNA-based oligonucleotide probe library (mercury LNA array ready to spot v.7.1, Exiqon), processed, and analyzed as described earlier (Schepeler et al., 2008) using TIGR spotfinder 2.23, TIGR MIDAS 2.19, and TIGR MEV 3.1 software. For gene expression microarray analysis, the RNA was labeled, hybridized to Human Exon 1.0 ST Arrays (Affymetrix, Santa Clara, CA, USA), and analyzed as described earlier (Thorsen et al., 2008).

Real-time RT–qPCR miRNA and mRNA expression analysis

TaqMan miRNA assays (Applied Biosystems, Foster City, CA, USA) were used for quantification of miRNA expression using ribosomal RNA RNU6B and RNU43 for normalization. For validation of target gene expression, TaqMan Gene expression assays (Applied Biosystems) were used for CLINT1, PPP3CA, and CBFB using GAPDH for normalization. miRNA and mRNA RT–qPCR was performed in triplicates and as described by the manufacturer using an ABI7500 PCR system (Applied Biosystems) and 7500 Fast system software for data analysis.

Detection of cell viability and death

Viability and death of the cells were analyzed by MTT (Sigma-Aldrich, St Louis, MO, USA) reduction and lactate dehydrogenase (LDH; cytotoxicity detection kit, Roche, Basel, Switzerland) assays essentially as described (Ostenfeld et al., 2005). The cell death mode was assessed by the morphology, nuclear condensation, and plasma membrane integrity of dying cells by phase contrast microscopy, and by staining the cells with 2.5 μg/ml Hoechst-33342 and 0.5 μmol/l SYTOX-Green (Molecular Probes, Eugene, OR, USA) using Zeiss Axiovert 40 CFL and Zeiss Axiovert 200M fluorescence microscopes. SYTOX-Green positivity was also examined on a FACSCalibur flow cytometer (Becton Dickinson, Heidelberg, Germany) measuring the mean fluorescence intensity in the FL1-H channel of 10 000 cells/sample and using CellQuestPro software.

Caspase activity measurement

The analysis of caspase 3/7 activity (DEVDase activity) was performed essentially as described (Ostenfeld et al., 2005). The kinetic cleavage of the substrate Ac-DEVD-AFC (Biomol, Plymouth Meeting, PA, USA) as measured by the liberation of AFC (excitation, 400 nm; emission, 489 nm) was measured in cell lysates using a Synergy HT multi-mode micro plate reader, (Biotek Industries Inc., Winooski, VT, USA).

Immunodetection of proteins

Immunodetection of proteins separated by SDS–PAGE and transferred to nitrocellulose membranes was performed with enhanced chemiluminescence western blotting agents (Amersham Biosciences, Fairfield, CT, USA). Primary antibodies against Bcl-2 (1:1000, clone 124, Dako, Glostrup, Denmark) and β-actin (1:10 000, Sigma-Aldrich) followed by appropriate peroxidase-conjugated secondary antibodies from Dako A/S.

Bioinformatic and statistical analysis

miR-145-regulated mRNAs were identified using scr-miR transfection as reference. The 22-target gene signature was deduced from the cell culture model in which miR-145 was overexpressed based on criteria listed in the Supplementary information. Presence of the 7mer AACUGGA miR-145 target site was examined in all mRNAs and analysis of target site enrichment was performed by both a binomial test using R (http://www.cran.r-project.org) and by performing 107 permutations of the data to estimate the false discovery rate. For analysis of gene networks affected by miR-145, Ingenuity Pathways analysis (IPA) was used (http://www.ingenuity.com). The ‘Global Functional Analysis’ identified the biological function significantly associated with the expression pattern.

Conflict of interest

The authors declare no conflict of interest.

References

  1. Baek D, Villen J, Shin C, Camargo FD, Gygi SP, Bartel DP . (2008). The impact of microRNAs on protein output. Nature 455: 64–71.

  2. Calin GA, Croce CM . (2006). MicroRNA signatures in human cancers. Nat Rev Cancer 6: 857–866.

  3. Calin GA, Sevignani C, Dumitru CD, Hyslop T, Noch E, Yendamuri S et al. (2004). Human microRNA genes are frequently located at fragile sites and genomic regions involved in cancers. Proc Natl Acad Sci USA 101: 2999–3004.

  4. Cordes KR, Sheehy NT, White MP, Berry EC, Morton SU, Muth AN et al. (2009). miR-145 and miR-143 regulate smooth muscle cell fate and plasticity. Nature 460: 705–710.

  5. Dyrskjot L, Ostenfeld MS, Bramsen JB, Silahtaroglu AN, Lamy P, Ramanathan R et al. (2009). Genomic profiling of microRNAs in bladder cancer: miR-129 is associated with poor outcome and promotes cell death in vitro. Cancer Res 69: 4851–4860.

  6. Dyrskjot L, Zieger K, Kruhoffer M, Thykjaer T, Jensen JL, Primdahl H et al. (2005). A molecular signature in superficial bladder carcinoma predicts clinical outcome. Clin Cancer Res 11: 4029–4036.

  7. Esquela-Kerscher A, Slack FJ . (2006). Oncomirs—microRNAs with a role in cancer. Nat Rev Cancer 6: 259–269.

  8. Hanahan D, Weinberg RA . (2000). The hallmarks of cancer. Cell 100: 57–70.

  9. Hart SM, Foroni L . (2002). Core binding factor genes and human leukemia. Haematologica 87: 1307–1323.

  10. Horrigan SK, Westbrook CA, Kim AH, Banerjee M, Stock W, Larson RA . (1996). Polymerase chain reaction-based diagnosis of del (5q) in acute myeloid leukemia and myelodysplastic syndrome identifies a minimal deletion interval. Blood 88: 2665–2670.

  11. Iorio MV, Ferracin M, Liu CG, Veronese A, Spizzo R, Sabbioni S et al. (2005). MicroRNA gene expression deregulation in human breast cancer. Cancer Res 65: 7065–7070.

  12. Kagoshima H, Nimmo R, Saad N, Tanaka J, Miwa Y, Mitani S et al. (2007). The C. elegans CBFbeta homologue BRO-1 interacts with the Runx factor, RNT-1, to promote stem cell proliferation and self-renewal. Development 134: 3905–3915.

  13. Kedde M, Strasser MJ, Boldajipour B, Oude Vrielink JA, Slanchev K, le Sage C et al. (2007). RNA-binding protein Dnd1 inhibits microRNA access to target mRNA. Cell 131: 1273–1286.

  14. Kim M, Park SY, Pai HS, Kim TH, Billiar TR, Seol DW . (2004). Hypoxia inhibits tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis by blocking Bax translocation. Cancer Res 64: 4078–4081.

  15. Kim R, Emi M, Tanabe K, Murakami S, Uchida Y, Arihiro K . (2006). Regulation and interplay of apoptotic and non-apoptotic cell death. J Pathol 208: 319–326.

  16. Kroemer G, Galluzzi L, Vandenabeele P, Abrams J, Alnemri ES, Baehrecke EH et al. (2008). Classification of cell death: recommendations of the Nomenclature Committee on Cell Death 2009. Cell Death Differ 16: 3–11.

  17. Lee Y, Ahn C, Han J, Choi H, Kim J, Yim J et al. (2003). The nuclear RNase III Drosha initiates microRNA processing. Nature 425: 415–419.

  18. Levine B, Yuan J . (2005). Autophagy in cell death: an innocent convict? J Clin Invest 115: 2679–2688.

  19. Litynska A, Przybylo M, Ksiazek D, Laidler P . (2000). Differences of alpha3beta1 integrin glycans from different human bladder cell lines. Acta Biochim Pol 47: 427–434.

  20. Michael MZ, SM OC, van Holst Pellekaan NG, Young GP, James RJ . (2003). Reduced accumulation of specific microRNAs in colorectal neoplasia. Mol Cancer Res 1: 882–891.

  21. Mitra AP, Birkhahn M, Cote RJ . (2007). p53 and retinoblastoma pathways in bladder cancer. World J Urol 25: 563–571.

  22. Nakajima G, Hayashi K, Xi Y, Kudo K, Uchida K, Takasaki K et al. (2006). Non-coding MicroRNAs hsa-let-7g and hsa-miR-181b are associated with chemoresponse to S-1 in colon cancer. Cancer Genomics Proteomics 3: 317–324.

  23. Nuovo GJ . (2008). In situ detection of precursor and mature microRNAs in paraffin embedded, formalin fixed tissues and cell preparations. Methods 44: 39–46.

  24. Orom UA, Nielsen FC, Lund AH . (2008). MicroRNA-10a binds the 5’UTR of ribosomal protein mRNAs and enhances their translation. Mol Cell 30: 460–471.

  25. Ostenfeld MS, Fehrenbacher N, Hoyer-Hansen M, Thomsen C, Farkas T, Jaattela M . (2005). Effective tumor cell death by sigma-2 receptor ligand siramesine involves lysosomal leakage and oxidative stress. Cancer Res 65: 8975–8983.

  26. Ow YL, Green DR, Hao Z, Mak TW . (2008). Cytochrome c: functions beyond respiration. Nat Rev Mol Cell Biol 9: 532–542.

  27. Porkka KP, Pfeiffer MJ, Waltering KK, Vessella RL, Tammela TL, Visakorpi T . (2007). MicroRNA expression profiling in prostate cancer. Cancer Res 67: 6130–6135.

  28. Rajewsky N . (2006). microRNA target predictions in animals. Nat Genet 38 (Suppl): S8–S13.

  29. Sachdeva M, Zhu S, Wu F, Wu H, Walia V, Kumar S et al. (2009). p53 represses c-Myc through induction of the tumor suppressor miR-145. Proc Natl Acad Sci USA 106: 3207–3212.

  30. Schepeler T, Reinert JT, Ostenfeld MS, Christensen LL, Silahtaroglu AN, Dyrskjot L et al. (2008). Diagnostic and prognostic microRNAs in stage II colon cancer. Cancer Res 68: 6416–6424.

  31. Selbach M, Schwanhausser B, Thierfelder N, Fang Z, Khanin R, Rajewsky N . (2008). Widespread changes in protein synthesis induced by microRNAs. Nature 455: 58–63.

  32. Sempere LF, Christensen M, Silahtaroglu A, Bak M, Heath CV, Schwartz G et al. (2007). Altered MicroRNA expression confined to specific epithelial cell subpopulations in breast cancer. Cancer Res 67: 11612–11620.

  33. Shi B, Sepp-Lorenzino L, Prisco M, Linsley P, deAngelis T, Baserga R . (2007). Micro RNA 145 targets the insulin receptor substrate-1 and inhibits the growth of colon cancer cells. J Biol Chem 282: 32582–32590.

  34. Tavazoie SF, Alarcon C, Oskarsson T, Padua D, Wang Q, Bos PD et al. (2008). Endogenous human microRNAs that suppress breast cancer metastasis. Nature 451: 147–152.

  35. Thorsen K, Sorensen KD, Brems-Eskildsen AS, Modin C, Gaustadnes M, Hein AM et al. (2008). Alternative splicing in colon, bladder, and prostate cancer identified by exon array analysis. Mol Cell Proteomics 7: 1214–1224.

  36. Vasudevan S, Tong Y, Steitz JA . (2007). Switching from repression to activation: microRNAs can up-regulate translation. Science 318: 1931–1934.

  37. Wang X, Tang S, Le SY, Lu R, Rader JS, Meyers C et al. (2008). Aberrant expression of oncogenic and tumor-suppressive microRNAs in cervical cancer is required for cancer cell growth. PLoS ONE 3: e2557.

  38. Yanaihara N, Caplen N, Bowman E, Seike M, Kumamoto K, Yi M et al. (2006). Unique microRNA molecular profiles in lung cancer diagnosis and prognosis. Cancer Cell 9: 189–198.

  39. Yang H, Kong W, He L, Zhao JJ, O’Donnell JD, Wang J et al. (2008). MicroRNA expression profiling in human ovarian cancer: miR-214 induces cell survival and cisplatin resistance by targeting PTEN. Cancer Res 68: 425–433.

  40. Yu L, Alva A, Su H, Dutt P, Freundt E, Welsh S et al. (2004). Regulation of an ATG7-beclin 1 program of autophagic cell death by caspase-8. Science 304: 1500–1502.

Download references

Acknowledgements

We thank Gitte Høj, Pamela Celis, Hanne Steen, Inge-Lis Thorsen, Gitte Stougård, and Conni Sørensen for technical assistance. We are grateful to M Jäättelä for providing the pCEP4 Bcl-2 vector construct and to Thomas B Hansen for methylation analysis software. We thank the staff at the Departments of Urology, Clinical Biochemistry, and Pathology at Aarhus University Hospital. This work was supported by the Ministry of Technology and Science, The John and Birthe Meyer Foundation, the Lundbeck Foundation, and the Danish Cancer Society, and the European Community's Seventh framework program (FP7/2007–2013) under grant agreement no 201663.

Author information

Correspondence to T F Ørntoft.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Ostenfeld, M., Bramsen, J., Lamy, P. et al. miR-145 induces caspase-dependent and -independent cell death in urothelial cancer cell lines with targeting of an expression signature present in Ta bladder tumors. Oncogene 29, 1073–1084 (2010) doi:10.1038/onc.2009.395

Download citation

Keywords

  • miRNA
  • bladder cancer
  • cell death
  • caspases
  • microarray

Further reading