Introduction

Uncontrolled cell proliferation due to aberrant regulation of cell cycle control can lead to the development of cancer. Cyclin dependent kinase inhibitors (CKIs) are the common inhibitors of cell cycle which consists of two families: INK4 family (consisting of p16INK4a, p15INK4b, p18INK4c and p19INK4d) and Cip/Kip family (p21CIP1/WAF1, p27Kip2 and p57Kip2) [1]. Human p57Kip2 gene is maternally expressed and paternally imprinted and is located on chromosome 11p15.5 [2], which is implicated in Beckwith-Wiedemann syndrome [3] and in sporadic cancers. p57Kip2 is also required for normal development as p57Kip2 null mice die at 2 weeks of age and show increased apoptosis and delayed differentiation during mouse development [4]. Because of its chromosomal location, imprinting status and functional activities, p57Kip2 is considered to be a candidate tumor suppressor gene. p57Kip2 overexpression in LNCaP prostate cancer cells resulted in the conversion of adenocarcinoma to a more differentiated squamous tumor in nude mice, with reduced cell proliferation and tumor invasion [5]. p57Kip2 is considered to be a tumor suppressor gene since it functions to block cell proliferation by inhibiting cell cycle progression, promotes apoptosis and cell differentiation, inhibits tissue invasion and metastasis and also inhibits angiogenesis [6]. Hence, cancer cells frequently down-regulate p57Kip2 in order to gain a proliferative advantage.

Loss of or reduced p57Kip2 expression occurs in carcinomas of the prostate, bladder, liver, pancreas, breast and others. The absence of p57Kip2 gene mutations in a wide variety of cancers suggests that other transcriptional or post-transcriptional mechanisms might be involved in its down-regulated protein expression. Inactivation of p57Kip2 gene due to promoter DNA methylation was observed in non-small cell lung cancer and in lymphoid malignancies of B-cells [7, 8]. Yang et al. found that p57Kip2 expression in breast cancer cells was repressed due to Polycomb protein EZH2-mediated H3K27me3 chromatin mark. Increased p57Kip2 degradation due to ubiquitylation by E3 ligase Skp1/ Cul1/ F-box (SCF complex) was observed in non-small cell lung carcinoma and hepatocellular carcinoma [9, 10].

Mature miRNAs are ~22 nucleotides long non-coding single-stranded RNAs, which upon binding to the 3’-UTR region of target mRNAs can result in mRNA cleavage, or translational repression. p57Kip2 has been reported to be targeted by miR-221/222 cluster in gastric carcinoma, ovarian cancer and hepatocellular carcinoma [11]. miR-92b and miR-25 have also been reported to down-regulate p57Kip2 expression in human embryonic stem cells and in gastric cancer respectively [12, 13]. Jin et al. demonstrated that expression of p57Kip2 is significantly decreased in human prostate cancer and overexpression of p57Kip2 in prostate cancer cells decreased cell proliferation and reduced invasiveness [5]. However, the mechanism behind p57Kip2 down-regulation in prostate cancer cells has not been investigated. Unexpectedly, we discovered that p57Kip2 is one of the novel downstream target genes of miR-21 in prostate cancer. We observed a very strong negative correlation between p57Kip2 and miR-21 expression in human prostate tumor samples and in CWR22, a human prostate cancer xenograft model. Our results for the first time show that miR-21 can down-regulate p57Kip2 mRNA and protein expression by targeting its coding region to attenuate its activity in prostate cancer cells. Thus, therapeutic approaches aimed at restoring p57Kip2 expression might be beneficial for prostate cancer prevention and therapy.

Findings and discussion

MicroRNA-21 targets p57Kip2 gene

Inactivation of p57Kip2 is commonly observed in cancers. MicroRNAs offer another layer of complexity to the regulation of its gene expression. Recently, microRNAs (miR-221/222, miR-25, miR-92b) were reported to downregulate the expression of p57Kip2 transcript [13, 14]. In our studies, we found a very significant p57Kip2 downregulation in human prostate tumor samples compared to the normal prostate cells in consistence with the published studies [5, 15]. However, the mechanism of p57Kip2 repression has not been well studied in prostate cancer. We along with others have shown a significant upregulation in miR-21 expression in prostate cancer [16, 17]. Surprisingly, we found a significant (P < 0.05) inverse correlation between miR-21 and p57Kip2 expression in human prostate cancer samples downloaded from the Cancer Genome Atlas (TCGA) database (Figure 1A). We also checked miR-21 expression in different prostate cancer cells and found that all the cells expressed miR-21 albeit at different levels. Cells with relatively higher miR-21 levels such as MDA-PCa-2b and 22Rv1 expressed lower levels of p57Kip2 when compared to PC-3 cells that expressed relatively low miR-21 and higher p57Kip2 (Figure 1B). Hence, we hypothesized that p57Kip2 may be a downstream target gene of miR-21 in prostate cancer. Regulation of target genes by binding of microRNA’s to the 3' untranslated regions (UTRs) of their mRNA’s has been well characterized. Hence, most of the bioinformatics tools available to predict miRNA target sites on mRNAs have been mainly focused on target sites within the 3'-UTRs of genes. A study by Hausser et al. suggest that miRNAs may combine targeting of coding regions and 3’ UTRs to flexibly tune their post-transcriptional regulatory effects [18]. Available microRNA prediction target algorithms did not predict p57Kip2 gene to be targeted by miR-21. However, scanning of human p57Kip2 gene for the miR-21 recognition site, revealed a 5-mer complementary binding site to the 2-7 seed region of mature miR-21 and a 3-mer sequence complementary to the 3’-target site (18-20 nucleotide) in the coding region of p57Kip2, which is from nucleotides 1054 to 1077 in accession # NM_00007 of GenBank as shown in Figure 1C. To experimentally test if miR-21 targets p57Kip2, we cloned this putative miR-21 recognition sequence of p57Kip2 in the 3’-UTR region of the luciferase gene in the pMIR reporter vector. Transfection of miR-21 mimic in PC-3 prostate cancer cells significantly (P < 0.01) reduced the luciferase activity (Figure 1D). However, mutation of the putative miR-21 binding nucleotides (Figure 1C and Additional file 1) abolished this effect. Similarly, luciferase activity was significantly (P < 0.01) increased when 22Rv1 prostate cancer cells, which were previously also shown to express relatively high endogenous miR-21 [16], were transfected with an anti-miR-21 to inhibit the endogenous miR-21 levels (Figure 1D), and the mutation of the miR-21 binding sites abrogated this effect.

Figure 1
figure 1

MicroRNA-21 targets p57Kip2gene. (A) miR-21 and p57Kip2 are inversely correlated in 37 matched normal (N) and tumor (T) samples downloaded from the TCGA database. r: Pearson product-moment correlation coefficient. ****P < 0.0001 with one-way analysis of variance. (B) miR-21 (left panel) and p57Kip2 (right panel) levels were measured in prostate cancer cells with real-time RT-PCR and western blotting respectively. *P < 0.05 with one-way analysis of variance test. (C) Sequence alignment between miR-21 seed sequence and a partial coding or mutated sequence of p57Kip2. (D) Luciferase assay was performed with miR-21 mimic or anti-miR-21 transfection in cells with co-transfection of pMIR-3’UTR- p57Kip2 or pMIR-3’UTR-mutated p57Kip2 and a ß-Gal expression construct. **P < 0.01 with one-way analysis of variance and Tukey-Kramer post hoc test.

Full size image

MicroRNA-21 downregulates p57Kip2 expression

To further validate these findings, we checked the effect of miR-21 on p57Kip2 mRNA and protein levels. Inhibition of endogenous miR-21 with an anti-miR-21 in 22Rv1 and MDA-PCa-2b cells increased the p57Kip2 mRNA levels (P < 0.05), whereas miR-21 mimic decreased (P < 0.001) the basal level of p57Kip2 in PC-3 cells (Figure 2A). Transfection of anti-miR-21 in 22Rv1 and MDA-PCa-2b cells increased the p57Kip2 protein expression and miR-21 mimic transfection in PC-3 decreased p57Kip2 expression (Figure 2B). We used a CWR22 xenograft mouse model to study the effect on microRNA-21 and p57Kip2 expression under androgen depletion with castration [19]. We have previously shown that castration led to down-regulation of miR-21 as androgen signaling was shown to stimulate miR-21 [16]. We harvested tumors for miR-21 and p57Kip2 expression analysis at 14 and 40 days post castration, respectively. The results showed a significant (P < 0.0001) decrease in miR-21 expression levels in the castrated mice when compared to control non-castrated mice (Figure 2C) as was observed in our previous study. Interestingly, we observed an increase in p57Kip2 expression at 14 and 40 days in the castrated mice showing an inverse correlation between miR-21 and p57Kip2 expression in vivo (Figure 2C) (P < 0.0001). We show that for the first time, castration-mediated miR-21 decreases correlate with an increase in p57Kip2 under in vivo conditions. Currently, it is not clear how androgen/AR signaling regulates p57Kip2 expression in prostate cancer cells. Our results suggest that androgen/AR signaling-mediated inhibition of p57Kip2 expression appears to be mediated by miR-21 in CWR22 human prostate cancer xenografts. Further studies are needed to understand whether p57Kip2 expression contributes to castration-induced regression of prostate cancer.

Figure 2
figure 2

MicroRNA-21 downregulates p57Kip2in prostate cancer cells. (A) 22Rv1 and MDA-PCa-2b cells were transfected with the inhibitor negative control and anti-miR-21 (50 nM) for 48 h. PC-3 cells were transfected with control siRNA and miR-21 mimic (20 nM) for 48 h. p57Kip2 gene expression was checked by real-time RT-PCR. *P < 0.05 and ***P < 0.001 with a Student’s t-test analysis. (B) Cell lysates were used for Western analysis to measure the level of p57Kip2 and GAPDH after anti-miR-21 inhibitor or miR-21 mimic transfection for 48 h. (C) p57Kip2 and miR-21 expressions were analyzed in CWR22 tumors after 14 and 40 days of castration. ****P < 0.0001 with one-way analysis of variance test.

Full size image

MicroRNA-21 abrogates p57Kip2-mediated functional responses in prostate cancer cells

p57Kip2 has been shown previously to act as a tumor suppressor gene by inhibiting cell migration and invasion [20]. Hence, we tested whether endogenous miR-21 functionally abrogated p57Kip2 mediated tumor suppressive responses in prostate cancer cells. When we inhibited the endogenous miR-21 in PC-3 cells with an anti-miR-21 inhibitor, we observed an induction in p57Kip2 expression (Figure 3A). However, co-transfection with a p57Kip2 small-interfering RNA with anti-miR-21 inhibitor abolished the p57Kip2 induction by anti-miR-21 (Figure 3A). We also found that p57Kip2 inhibition significantly increased the cell migration in both PC-3 and 22Rv1 cells (Figure 3B). Transfection of cells with an anti-miR-21 inhibitor decreased the cell migration, although the results were not statistically significant. Interestingly, we found that the co-transfection of cells with anti-miR-21 and p57Kip2 siRNA, significantly increased cell migration when compared to anti-miR-21 treatment alone (Figure 3B) (P < 0.05 and P < 0.01). Similarly, while knocking down of endogenous miR-21 significantly reduced anchorage-independent growth, knocking down p57Kip2 reversed the inhibitory effect of the anti-miR-21 inhibitor on the anchorage-independent cell growth (Figure 3C) (P < 0.01 and P < 0.001). These results show that the functional effects of miR-21 in prostate cancer cells could be mediated partly by a downregulation in p57Kip2 expression.

Figure 3
figure 3

MicroRNA-21 attenuates p57Kip2mediated functional responses in prostate cancer cells. Cells were transfected with 50 nM of control siRNA, anti-miR-21 or p57Kip2 siRNA either alone or in combination for 48 h. Cells were trypsinized and replated for Western analysis, cell migration or soft agar assay. (A) Western analysis in PC-3 cells for p57Kip2 and GAPDH was performed. (B) Number of cells migrated were quantified. *P < 0.05 and **P < 0.01 with one-way analysis of variance test. (C) After 14 days, colonies were stained and quantified in PC-3 and 22Rv1 cells for soft-agar colony formation assay. *P < 0.05, **P < 0.01 and ***P < 0.001 with one-way analysis of variance test.

Full size image

In summary, we discovered p57Kip2 to be a novel target gene of microRNA-21 in prostate cancer. Our findings provide a novel mechanism of p57Kip2 downregulation in prostate cancer. These findings warrant further research to test if p57Kip2 is also a novel target gene of miR-21 in other cancer types as well. In our previous study, we found miR-21 to be an oncogenic regulator in prostate cancer by targeting the tumor suppressive effects of TGF-beta signaling pathway in cancer cells [16]. Given, the regulation of p57Kip2 by microRNA-21 in the current study, we provide a strong rationale to perform preclinical testing of microRNA-21 inhibitor as a novel therapeutic drug for prostate cancer.