Introduction

The current 5-year survival rate of patients with localized prostate cancer (PCa) is almost 100%, which is attributed to efficacy of treatments of early-stage disease.1 Although most patients with PCa are initially responsive to androgen-deprivation therapy, their cancers eventually become resistant to this mode of treatment, and they progress to castration-resistant PCa.1, 2 Unfortunately, most clinical trials for advanced PCa have shown limited benefits, eventually resulting in disease progression and metastasis.1, 2 Distant metastasis of advanced PCa is common and leads to significant morbidity and mortality. Therefore, we have proposed that current therapies for advanced PCa could be improved by using new genomic approaches to better understand the molecular mechanisms of metastasis and disease progression.

MicroRNAs (miRNAs) are endogenous small non-coding RNA molecules (19–22 bases in length) that regulate the expression of protein-coding or protein non-coding genes in a sequence-specific manner.3, 4 There is substantial evidence that miRNAs can be oncogenic or act to suppress tumors. Dysregulated expression of oncogenic miRNAs can disrupt the normally controlled RNA networks present in normal cells and enhance the development of cancer.5, 6, 7, 8 Identification of aberrantly expressed miRNAs in cancer cells provides important new information regarding the initiation, progression and metastasis of cancer cells.

We have identified antitumor miRNAs and miRNA-mediated oncogenic genes in PCa cells by using our PCa miRNA expression signatures.9, 10 Our recent studies showed that six miRNAs (miR-26a, miR-26b, miR-29a, miR-29b, miR-29c and miR-218) were downregulated in PCa tissues and restoration of these miRNAs markedly suppressed cancer cell aggressiveness.11, 12, 13 Our data suggested that these six miRNAs act as antimetastatic miRNAs in PCa cells. Therefore, we hypothesized that the genes regulated by these miRNAs significantly contributed to PCa metastasis. The aim of this study was to identify the genes targeted by these six antitumor miRNAs and to investigate their functional roles in migration and invasion in PCa cells.

Present study identified that a total of 35 putative metastasis-promoting genes were targeted by the 6 abovementioned miRNAs. Here we focused on the lysyl oxidase-like 2 (LOXL2) gene, a member of the lysyl oxidase (LOX) family.14 The function of LOXL2 is to promote crosslinking of collagen and elastin in the extracellular matrix (ECM).15, 16, 17 Past studies showed that LOXL2 contributed to the regulation of extracellular and intracellular cell signaling pathways.15, 16, 17 Our recent studies demonstrated that LOXL2 was controlled by several tumor-suppressive miRNAs and promoted cancer cell aggressiveness in renal cell carcinoma, head and neck cancer and lung cancer.18, 19, 20 Furthermore, several studies demonstrated that upregulation of LOXL2 occurred in many types of cancers and its expression contributed to cancer cell metastasis.21, 22, 23 Our present data show that insights into the molecular mechanisms of PCa pathogenesis can be revealed by identification of antitumor miRNA-regulated RNA networks.

Materials and methods

PCa cell lines and RNA extraction

Two human PCa cell lines (PC3 and PC3M) was obtained from ATCC (Manassas, VA, USA). The maintenance of PCa cells and RNA extraction procedures were as described previously.11, 24

Quantitative real-time reverse transcription–PCR (qRT–PCR)

TaqMan probes and primers of LOXL2 (P/N: Hs00158757_ml; Applied Biosystems, Foster City, CA, USA) and GAPDH (P/N: Hs02758991_gl; Applied Biosystems) was used as an internal control. PCR quantification method was as described previously.13, 25, 26

Transfection with miRNA mimic and small-interfering RNA (siRNA)

The following miRNA mimics were used in the present analysis: Ambion (Austin, TX, USA) Pre-miR miRNA precursor for hsa-miR-26a-5p (product ID: PM10249), hsa-miR-26b-5p (product ID: PM12899), hsa-miR-29a-3p (product ID: MC12499), hsa-miR-29b-3p (product ID: MC10103), hsa-miR-29c-3p (product ID: MC10518) and hsa-miR-218 (product ID: AM17100). Following siRNAs were used in this study: si-LOXL2 (P/N: HSS180848, HSS106125; Invitrogen, Carlsbad, CA, USA) and negative control miRNA/siRNA (P/N: AM17111; Applied Biosystems). The procedures of transfection of miRNAs and siRNAs were as described previously.12, 13

Selection of putative miR-26a / b-, miR-29a / b / c- and miR-218-regulated targets in PCa cells

To identify putative targets of these miRNAs, we carried out a combination of in silico database analysis and comprehensive gene expression analysis, as described in previous studies.12, 13, 25 Our strategy for identification of putative targets is shown in Figure 1.

Figure 1
figure 1

Workflow of the strategy for selection of putative target genes regulated by tumor-suppressive miRNAs (miR-26a/b, miR-29s and miR-218) in PCa. We screened candidate targets using the TargetScan database. In total, 334 genes were identified as putative targets containing binding sites for miR-26a/b, miR-29s and miR-218. Among these, 35 genes were downregulated in miR-26a-, miR-29a- and miR-218-transfected PC3 cells (log2 ratio <0). The sequences of miR-26a/b, miR-29 family and miR-218 miRNAs. Seed sequences are shown by red letters.

Full size image

Immunohistochemistry and western blotting

A tissue microarray containing PCa, prostatic intraepithelial neoplasias (PINs) and prostatic hyperplastic tissues was obtained from Provitro (Berlin, Germany) (Cat. no. 401 2209, Lot no. 146.1 P020212, 26–46). The information about these tissues can be found at http://www.provitro.com/fileadmin/provitro-data/TMA/4012209.pdf. Tissue immunostaining of LOXL2 and scoring method were as described previously.24, 25

The procedures of LOXL2 western blotting were as described previously.12, 13, 25

Cell proliferation, migration and invasion assays in PCa cells

To know the functional significance of LOXL2 in PCa cells, we investigated the knockdown effects of LOXL2 on cell proliferation, migration and invasion assays using si-LOXL2-transfected PC3 and PC3M cells. These assays were as described previously.12, 13, 25

Plasmid construction and dual-luciferase reporter assays

Partial wild-type sequences of the LOXL2 3′ untranslated region (3′-UTR) or those with deleted sequences of miRNAs (miR-26a/b-, miR-29a/b/c- and miR-218-binding sites) were inserted in the hRluc gene in the psiCHECK-2 vector (C8021; Promega, Madison, WI, USA). The procedure for the dual-luciferase reporter assay was as described previously.12, 13, 25

Statistical analysis

Relationships between two or three variables and numerical values were investigated using the Mann–Whitney U-test or Bonferroni adjusted Mann–Whitney U-test. Expert StatView software, version 4 (SAS Institute Inc., Cary, NC, USA), was used in these analyses.

Results

Identification of candidate targets regulated by antitumor miRNAs in PCa cells

To identify target genes of the six miRNAs, we carried out a combination of in silico database analysis and genome-wide gene expression analysis. Seed sequences of miR-26a and miR-26b are identical, and seed sequences of miR-29a, miR-29b and miR-29c are identical (Figure 1). First, we screened putative target genes using the TargetScan database Release 6.2 (http://www.targetscan.org/) and selected 334 genes that had putative binding sequences for these miRNAs in their 3′-UTRs. Next we narrowed down the list of genes by oligomicroarray analysis using PC3 cells and found that 35 genes were downregulated (log2 ratio <0) following transfection with miR-26a or miR-29a or miR-218 as compared with expression levels in miR-control-transfected cells (Table 1). Expression data were deposited in the Gene Expression Omnibus database (accession numbers, GSE56243 and GSE77790).

Table 1 Putative target genes regulated by miR-26a, miR-29a and miR-218 in PCa cells
Full size table

We investigated the regulation of hyaluronan synthase 3 (HAS3), LOXL2 and LIM and SH3 protein 1 (LASP1) by these antitumor miRNAs. The data showed that miR-26a and miR-26b slightly suppressed the expression of HAS3 and LASP1 in PCa cells (data not shown). Therefore, we excluded HAS3 and LASP1 genes in this analysis. In this study, we focused on LOXL2 because many reports showed that this gene had a pivotal role in cancer metastasis. Furthermore, we previously reported that miR-29a/b/c acted as antitumor miRNAs via targeting of LOXL2 in lung squamous cell carcinoma and renal cell carcinoma.18, 20

Direct regulation of LOXL2 by antitumor miRNAs in PCa cells

First, qRT–PCR and western blotting were carried out in PC3 and PC3M cells transfected with miR-26a, miR-26b, miR-29a, miR-29b, miR-29c or miR-218 to investigate whether LOXL2 expression was downregulated by restoration of these miRNAs. The expression levels of LOXL2 mRNA and protein were markedly suppressed in transfected cells with these miRNAs (Figures 2a, b, 3a). Next we performed luciferase reporter assays in PC3 cells to determine whether LOXL2 mRNA had actual target sites for these miRNAs. We used vectors encoding either a partial wild-type sequence or a sequence in which the miRNA-binding sequence was deleted from LOXL2 mRNA. The luminescence intensity was reduced by co-transfection with miR-26a, miR-26b, miR-29a, miR-29b, miR-29c, miR-218 and the vector carrying the wild-type LOXL2 3′-UTR in PC3 cells (Figures 2c, 3c and 4c).

Figure 2
figure 2

Direct regulation of LOXL2 by miR-26a and miR-26b in PCa cell lines. (a) LOXL2 mRNA expression 72 h after transfection with miR-26a or miR-26b. GAPDH expression was used for normalization. *P<0.0001. (b) LOXL2 protein expression 72 h after transfection with miR-26a/b. GAPDH was used as a loading control. (c) The miR-26a- and miR-26b-binding site in the 3′-UTR of LOXL2 mRNA. Luciferase reporter assays were performed using vectors that included (WT) or lacked (DEL) the wild-type sequences of the putative miR-26a and miR-26b target site. *P<0.0001.

Full size image
Figure 3
figure 3

Direct regulation of LOXL2 by miR-29a, miR-29b and miR-29c in PCa cell lines. (a) LOXL2 mRNA expression 72 h after transfection with miR-29s. GAPDH expression was used for normalization. *P<0.0001. (b) LOXL2 protein expression 72 h after transfection with miR-29s. GAPDH was used as a loading control. (c) The miR-29s-binding site in the 3′-UTR of LOXL2 mRNA. Luciferase reporter assays were performed using vectors that included (WT) or lacked (DEL) the wild-type sequences of the putative miR-29s target site. *P<0.0001.

Full size image
Figure 4
figure 4

Direct regulation of LOXL2 by miR-218 in PCa cell lines. (a) LOXL2 mRNA expression 72 h after transfection with miR-218. GAPDH expression was used for normalization. *P<0.0001. **P=0.0002. (b) LOXL2 protein expression 48 h after transfection with miR-218. GAPDH was used as a loading control. (c) The miR-218-binding site in the 3′-UTR of LOXL2 mRNA. Luciferase reporter assays were performed using vectors that included (WT) or lacked (DEL) the wild-type sequences of the putative miR-218 target site. Renilla luciferase assays were normalized to firefly luciferase values. *P<0.0001.

Full size image

Effects of knockdown of LOXL2 on cell proliferation, migration and invasion in PCa cell lines

To investigate the cancer cell-promoting role of LOXL2, we performed loss-of-function assays using si-LOXL2 transfectants. Knockdown efficiency of si-LOXL2 was evaluated in PC3 and PC3M transfectant cells. Both mRNA and protein expression levels of LOXL2 were downregulated in PC3 and PC3M cells (Figures 5a and b). XTT cell proliferation, wound-healing and invasion assays demonstrated that cell proliferation, migration and invasion activities were significantly suppressed in si-LOXL2-transfectant cells (Figures 5c–e).

Figure 5
figure 5

Effects of si-LOXL2 transfection on cell proliferation, migration and invasion in PCa cell lines. (a) LOXL2 mRNA expression levels were measured by qRT–PCR 72 h after transfection with 10 nm si-LOXL2. GAPDH was used for normalization. *P<0.0001. (b) LOXL2 protein expression 72 h after transfection with 10 nm si-LOXL2. GAPDH was used as a loading control. (c) Cell proliferation was determined by XTT assay 72 h after transfection with 10 nm si-LOXL2. *P<0.0001. (d) Cell migration activity was determined by migration assay 48 h after transfection with 10 nm si-LOXL2. *P<0.0001. (e) Cell invasion activity was determined by Matrigel invasion assay 48 h after transfection with 10 nm si-LOXL2. *P<0.0001.

Full size image

Immunostaining of LOXL2 in PCa tissues

We validated the expression levels of LOXL2 in PCa tissues using immunostaining. We used a tissue microarray containing 51 PCa, 10 PINs and 10 prostatic hyperplastic specimens (Table 2). Upregulation of LOXL2 protein was confirmed in the PCa tissues compared with noncancerous tissues (Figure 6).

Table 2 Clinical characteristics and IHC score of LOXL2 in tissue microarray
Full size table
Figure 6
figure 6

Immunohistochemical staining of LOXL2 in PCa clinical specimens. LOXL2 was expressed more strongly in several cancer lesions, weakly stained in PIN lesions than in normal tissues. (a) Prostate cancer, pT4N0, Grade3a, Gleason score 3+4; (b) Prostate cancer, pT2bN0, Grade3a, Gleason score 4+3; (c) PIN; (d) normal prostate tissue. (e) Quantification of LOXL2 expression. Expression of LOXL2 was upregulated in PCa and PIN specimens compared with normal prostate tissues.

Full size image

Discussion

Aberrantly expressed miRNAs might disturb normally regulated RNA networks and contribute to cancer cell pathology. Strategies to identify aberrant expression of miRNA-mediated RNA networks are being developed as a new direction in cancer research in the postgenome sequencing era. Based on the miRNA expression signature of PCa cells, we have continued the identification of tumor-suppressive miRNAs and their regulated novel PCa metastatic pathways.10, 11, 12, 13, 24, 25, 26, 27 Our recent studies showed that restoration of antitumor miRNAs (miR-26a, miR-26b, miR-29a, miR-29b, miR-29c and miR-218) markedly suppressed PCa cell aggressiveness via targeting of La-related protein 1 (LARP1), laminin γ1 (LAMC1) and LASP1, respectively.11, 12, 13 Moreover, these miRNA-targeted genes were overexpressed in PCa tissues and expression of these genes enhanced cancer cell migration and invasion.11, 12, 13 LASP1 encodes a LIM motif at its N-terminus and an src homology 3 (SH3) domain at its C-terminus. In LASP1, SH3 domain functions in protein–protein interactions28, 29 and LASP1 acts as an actin-binding protein and overexpression of LASP1 enhances cancer cell proliferation and invasion.13, 28, 30 Our present data showed that LASP1 is a putative candidate of these antitumor miRNA regulation, suggesting that LASP1 is a key molecule for cancer cell aggressiveness.

The miR-26 family is comprised of three subtypes in the human genome: miR-26a-1 (chromosome 3p22.2), miR-26a-2 (12q14.1), and miR-26b (2q35). The seed sequences of these miRNAs are identical, suggesting that the miR-26 family might regulate the same genes in human cells (miRBase: release 21; http://www.mirbase.org/). Downregulation of the miR-26 family and its antitumor effects have been reported in several cancers, such as bladder cancer, breast cancer, hepatocellular carcinoma, oral cancer and PCa.31, 32, 33, 34 Our past study showed that downregulation of these miRNAs enhanced cancer cell migration and invasion in oral cancer through direct regulation of TMEM184B.35 In PCa, EZH2, a histone-lysine N-methyltransferase enzyme, was directly regulated by miR-26a and miR-26b.36 Overexpression of EZH2 is observed in several cancers and contributes to cancer aggressiveness.36, 37

The miR-29 family consists of four miRNAs: miR-29a, miR-29b-1, miR-29b-2, and miR-29c. These miRNAs form cluster miRNA on two human chromosome loci: miR-29b-1 and miR-29a in 7q32 and miR-29b-2 and miR-29c in 1q32 (miRBase: release 21; http://www.mirbase.org/). Our miRNA expression signatures showed that all member of miR-29 family were downregulated in several cancers and that restoration of these miRNAs markedly inhibited cancer cell aggressiveness via targeting of ECM–integrin pathways.12, 38 Antitumor roles of the miR-29 family were demonstrated in several cancers.12, 18, 20, 38, 39 Interestingly, recent studies demonstrated molecular mechanisms of silencing of miR-29 family expression in cancer cells.40 The genome structure of miR-29b-1/miR-29a promoter region contains two E-box (MYC-binding) sites and four nuclear factor-κB-binding sites such that overexpressed Myc and nuclear factor-κB inhibited the expression of miR-29b-1/miR-29a at the transcriptional level.40 Overexpression of Myc was frequently observed in advanced PCa,41 and this phenomenon might enhance PCa cell progression and metastasis.

The miRNA database indicates that the miR-218 family is distributed between two human chromosomal loci: miR-218-1 at 4p15.31 and miR-218-2 at 5q35.1. Similar to miR-26a, miR-26b and miR-29 family, antitumor function of miR-218 have been described in many types of cancers.13, 42, 43, 44 Our past studies demonstrated that loss of miR-218 enhanced cancer cell migration and invasion through dysregulation of genes involved in the focal adhesion pathway, including CAV2, LAMA3 and LASP1.13, 42, 44 Two miRNAs, miR-218-1 and miR-218-2, are located on the introns of SLIT2 and SLIT3 genes, respectively, and control of the expression depends on the same promoter in their host genes.45 Several reports showed that the promoter regions of these genes were frequently methylated in cancer cell lines and clinical specimens.45, 46 Thus hypermethylation of their promoter regions caused silencing of miR-218-1 and miR-218-2 expression in cancer cells.

The highly invasive properties of PCa cells cause distant metastasis in patients with PCa, and metastasis is the primary reason for the high mortality of advanced PCa.1, 2 Studying the non-coding RNA networks could reveal the molecular mechanisms underlying metastatic pathways and facilitate the development of novel therapies to block progression of the disease. In this study, we hypothesized that several antitumor miRNAs (the miR-26a/b, the miR-29 family and miR-218) coordinately regulate genes that have key roles in PCa metastasis.

Here we focused on the LOXL2 gene as a putative regulatory target of these antitumor miRNAs in PCa cells. Recently, we showed that LOXL2 was directly regulated by miR-29a, miR-29b and miR-29c and its expression promoted cancer cell aggressiveness in renal cell carcinoma, non-small cell lung cancer and head and neck squamous cell carcinoma.18, 19, 20 Present data demonstrated that six antitumor miRNAs (miR-26a, miR-26b, miR-29a, miR-29b, miR-29c and miR-218) directly regulated LOXL2 in PCa cells. Moreover, overexpression of LOXL2 was confirmed in PCa tissues and knockdown of LOXL2 markedly impaired cancer cell aggressiveness. Notably, we showed that LOXL2 regulation by these miRNAs was also observed in head and neck squamous cell carcinoma cells.20 To identify regulation of LOXL2 by these miRNAs is needed to elucidate the metastatic roles of antitumor miRNAs in cancer.

The LOX protein family is comprised of five proteins (LOX and LOXL1–LOXL4). Their primary functions appear to be covalent crosslinking of collagen to elastin in the ECM.14, 15, 16, 17 Overexpression of the LOX family was observed in several cancers,18, 19, 20 indicating that dysregulated expression of the LOX family enhances ECM deposition and subsequent tissue stiffness. Overexpression of ECM component proteins is frequently observed in many types of cancer tissues. This aberration promotes cancer cell aggressiveness by dysregulation of cell adhesion and ECM remodeling.14 Thus aberrant expression of the LOX family is a trigger of malignant transformation of cancer cells through ECM dysregulation. Several studies have indicated that high expression of LOXL2 was correlated with poor prognosis in patients with gastric, breast, lung and laryngeal cancers.19, 20, 21, 22 To investigate the clinical features of PCa patients and LOXL2 expression is an important theme. A large scale of cohort study will be necessary in future. Interestingly, transcriptional control of LOXL2 was regulated by hypoxia-inducible factor-1. LOXL2 directly interacts with transcriptional factor SNAIL1 in the nucleus and repressed the expression of E-cadherin.47, 48 Thus hypoxic conditions induced LOXL2 expression, and the accumulating LOXL2 enhanced the aberrant activation of epithelial–mesenchymal transition signaling in cancer cells.

In conclusion, direct regulation of LOXL2 by antitumor miRNAs (miR-26a, miR-26b, miR-29a, miR-29b, miR-29c and m iR-218) was observed in PCa cells. Overexpression of LOXL2 was validated in PCa tissues and aberrantly expressed LOXL2 enhanced PCa cell aggressiveness. Understanding of novel RNA networks regulated by the antitumor miRNAs may lead to a better understanding of PCa metastasis.