MicroRNAs 221/222 and Genistein-Mediated Regulation of ARHI Tumor Suppressor Gene in Prostate Cancer

Prostate cancer is the most commonly diagnosed cancer among males in United States. It is estimated that approximately 192,280 new cases will be reported with 27,360 deaths in 2009. Although the prostate cancer death rate has declined with the improvement of detection and diagnosis since the 1990s, it remains the second most common cancer leading to death in men in the United States. So far, the molecular mechanisms involved in prostate cancer are not fully understood.

ARHI is a maternally imprinted human tumor suppressor gene that maps to chromosome 1p31 and encodes a 26-kDa small G protein with 60% homology to rap and ras (1). ARHI has been reported to be expressed in several normal tissues including ovarian (1), breast (2, 3), liver (4), thyroid, (5) and pancreas (6). However, its expression is lost or downregulated in their respective cancer tissues. Several mechanisms for this loss have been studied, for example, loss of heterozygosity contributes up to 40% of ARHI gene loss in ovarian and breast cancers (2, 7). ARHI is also downregulated by epigenetic mechanisms like DNA methylation, histone methylation, and histone acetalyation (8, 9). It has also been reported that ARHI is downregulated by binding of transcription factor E2F to the ARHI promoter region (10, 11). This kind of negative regulation can be reversed by TSA treatment, E2F binding site mutation and E2F siRNA (11).

Overexpression of ARHI in ovarian and breast cancer cells can inhibit cell growth, decrease cell invasiveness and cause caspase-independent apoptosis (12). Meanwhile, expression of ARHI at physiologic levels induces autophagy rather than apoptosis (13). Furthermore, low expression of ARHI has been closely related to shorter progression-free survival in pancreatic cancer (14).

However, little is known about ARHI expression in prostate cancer and whether its misregulation contributes to prostate cancer tumorigenesis. In this study, we investigated ARHI expression levels in prostate cancer cells and prostate cancer tissues. We also explored the functional role of ARHI in the pathogenesis of prostate cancer and for the first time show a new mechanism of ARHI regulation by miRNAs and genistein, a dietary isoflavone from soy.

Immortalized normal human prostate epithelial cells RWPE-1 and several prostate cancer derived cell lines (PC-3, LNCap, Du145) were employed in this study. RWPE-1 cells were cultured in Keratinocyte Serum-free Medium (K-SFM) (GIBCO Kit Catalog Number 17005–042) which was supplied with bovine pituitary extract and human recombinant epidermal growth factor Prostate cancer derived cell lines were cultured in RPMI media (UCSF facility) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (UCSF facility).

Prostate cancer tissue arrays were obtained from US Biomax, which included 95 cores, 48 cases with 36 cancers, 8 matched normal adjacent tissues, and 4 cases of bone metastasis.

Total RNA was extracted from cultured cells using an RNeasy mini kit or miRNeasy Mini Kit (for collection of small RNAs; Qiagen). Gene expression was measured by real-time quantitative PCR (RT-QPCR) using an Applied Biosystems 7500 Fast Sequence Detection System and gene-specific TaqMan assay kits. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and RNU48 were used as endogenous controls to normalized expression data. The thermal cycling conditions were according to the TaqMan Fast Universal PCR protocol. Each sample was analyzed in quadruplicate and each result was concluded from 3 independent experiments.

To study the expression profile of genes in a particular pathway, pathway-specific PCR arrays (cell cycle array, Apoptosis Array; SABioscience) were used according to the manufacturer's instructions. Data were analyzed by the instrument's software. Dysregulated genes were further validated by TaqMan gene expression assays (Applied Biosystems)

Tissue Scan q-PCR array (prostate cancer panel I; Origene) contains 48 prostate samples (7 adjacent normal, 10 tumor, 31 lesions). Only the data from clearly defined adjacent normal and tumor tissues were compared with each other for ARHI gene expression. The manufacturer's protocol was followed.

To prepare single-cell suspensions for flow cytometry, cell culture plates were incubated with 0.25% trypsin-EDTA (Invitrogen) at 37°C for 5 minutes and released by pipetting. The cell pellet was washed 3 times with ice-cold PBS and resuspended in 1 mL of nuclear stain DAPI (Beckman Coulter) for cell cycle analysis. Stained cells in a volume of 1 mL were analyzed by Flow Cytometer (Cell Lab Quanta SC). For apoptosis analysis, cells were double stained with viability dye 7-ADD and Annexin-V-FITC using Annexin V-FITC/7-AAD Kit (Beckman Coulter) and then analyzed.

PC-3 cells were plated in 96-well microplates at a density of 5 × 10³ cells per well 24 hours before transfection. Cell proliferation was determined by CellTiter 96 Aqueous One Solution Cell Proliferation Assay Kit (Promega) at 24, 48, and 72 hours after transfection. Absorbance was recorded at 490 nm with a kinetic microplate reader (Spectra MAX 190; Molecular Devices Co.). Experiments were performed 3 times in quadruplicate.

For invasion assay, we used BD Matrigel coated chambers (BD Bioscience) according to the manufacture's protocol. Briefly, 48 hours after cells were transfected with ARHI plasmid or control plasmid; cells were counted and plated on Matrigel inserts or control inserts at an equal number (1 × 10⁵cells/mL) in serum-free medium and allowed to grow for 15 hours at 37°C. Cells migrating through the membrane were fixed, stained with a Hema 3 kit (Fisher Chemicals) and counted manually.

Whole cell extracts were prepared from RIPA buffer (Thermal Scientific, Pierce). Protein lysate (50 μg) was electrophoresed in 12% polyacrylamide gels and were transferred to PVDF membrane at 75 mA overnight at 4°C. Next day, the PVDF membrane was blocked and incubated with 1:100 polyclonal goat anti-ARHI IgG (sc-30321; Santa Cruz Biotechnology) overnight at 4°C. The secondary antibody was HRP-conjugated horse anti-goat IgG (Cell signaling) diluted 1:2,500 in fresh blocking solution. ARHI was visualized using the Super Signal West Pico Chemiluminescence Detection Kit (Pierce). Membranes were exposed to X-OMAT AR film (Kodak). Antibody of GADD45A (ab54740), CASP1 (ab54932), CD70 (ab96323), P21 (ab16767), and HRK (ab84822) were obtained from Abcam.

Immunohistochemistry was done on formalin-fixed, paraffin-embedded prostate cancer tissue arrays. The slice was deparafinized and antigen retrieval was carried out by microwaving the slices in 10 mmol/L sodium citrate buffer. The staining kit was a Goat ImmunoCruz Staining System (sc-2053; Santa Cruz Biotechnology). The primary antibody was the same for Western blots. All procedures strictly followed the manufacture's protocol. The intensity of the staining was graded by 2 individuals from “−” as no staining to “++++” as strong staining. Figures were showing at 200× magnification in both adjacent normal and cancer tissues.

Genomic DNA from RWPE-1, PC-3, DU145, LNCap cells were treated with bisulfate by using an EpiTect Bisulfite kit (Qiagen) as per the manufacturer's instructions. Treated DNAs were analyzed for methylation status by using the specific methylation sequence. The specific primers were designed for the CpG island of the ARHI gene as followings:

• CpG island I

•   Forward: ATAGGTAAGGGAGAAAGAAGTTAGA;

•   Reverse: AATACTACTATCCTAACAAAACCCTC.

• CpG island II

•   Forward: TGTTTGAAGTTTATTTTAATAGGTGAGAA;

•   Reverse: ACTAAACACCCCAAAAACTCACTC.

• CpG island III

•   Forward: GGGGTATTTGAAAATAATAAGTGTTAGATA

•   Reverse: TCCAAAATACAATTACAAAACCTACC

The sequencing was done by an outside laboratory (MCLAB).

According to the miRDB analysis, oligonucleotides containing human ARHI 3′-UTR target sequence were annealed and cloned into the pMIR-REPORTER Luciferase plasmid (Applied Biosystem) following the manufacture's protocol to generate vector pMIR-ARHI-221/222.

• ARHI-3′UTR for miR-221

•   Forward: CTAGTATTGGCATTTGAAGAGGAGAACAA-AATGTAGCATGATGTATTA

•   Reverse: ATAACCGTAAACTTCTCCTCTTGTTTTACAT-CGTACTACATAATTCGA

• ARHI-3′UTR for miR-222:

•   Forward: CTAGTTGGCATTTGAAGAGGAGAACAAAA-TGTAGCATGATGTATTA

•   Reverse: AACCGTAAACTTCTCCTCTTGTTTTACATC-GTACTACATAATTCGA

Empty vector pMIR without the inserts was using as a negative control. pMIR-REPOTR β-gal control plasmid was used for transfection normalization. PC-3 cells were cultured in 24 well plates and transfected with 800 ng of pMIR-ARHI or pMIR together with 100 ng of pMIR-β-gal and 20 pmol of miR-221/222(Applied Biosystems). Lipofectamine 2000 and OPTI-MEMI reduced serum medium (Invitrogen) were used for transfection. Forty-eight hours after transfection, luciferase and β-gal activity were measured using the Dual-Light System (Applied Biosystems).

MicroRNA inhibitors were specifically designed by Applied Biosystems for targeting the respective miRNAs (miR-221 inhibitor catalog # AM 10337, miR-222 inhibitor catalog no. PM11376).

Statistical analysis was performed with Statview 5.0 for Windows as needed. P values less than 0.05 were regarded as statistically significant and are represented by an asterisk on the bars in the figures.

To evaluate ARHI expression levels in prostate cancer cell lines, ARHI mRNA level and protein levels were measured in immortalized prostate cell line RWPE-1 and 2 prostate cancer cell lines: PC-3, LNCap by real-time PCR and Western blot. We found that both ARHI mRNA and protein levels were dramatically downregulated in prostate cancer cell lines compared with RWPE-1 cells (Fig. 1A).

We also examined ARHI mRNA levels in prostate tissues by Q-PCR that consisted of 7 normal adjacent prostate tissues, 10 stage II–III prostate cancers, and 31 prostate samples with various levels of lesions. Compared with the normal adjacent controls, over 80% of the marked tumor samples had decreased expression of ARHI (Fig. 1B); suggesting that ARHI is frequently downregulated in prostate cancers at the mRNA level.

To further confirm whether expression of ARHI was also reduced in prostate cancer tissues at the protein level, a prostate cancer tissue array (36 cancers, 8 of which were matched adjacent normal tissue and 4 cases of bone metastasis) was immunostained with antiARHI antibody. ARHI showed expression in both epithelial cytoplasm and stoma, but the staining intensity was much higher in the epithelial cells. ARHI protein levels were observed to be downregulated in over 70% of the prostate cancer tissues as shown in representative samples of ARHI immunohistochemistry (Fig. 1C).

To understand the function of ARHI in prostate cancer, we transiently transfected PC-3 cells with a TrueORF pCMV6 expression vector containing the ARHI ORF. We chose PC-3 cells due to its low endogenous expression of ARHI. The empty vector of pCMV6 was used as a control. Real-time PCR showed a 20,000-fold overexpression of ARHI mRNA compared with the controls (Fig. 2A). We also performed Western analysis for ARHI protein levels to confirm the successful transfection (Fig. 2B). Forty-eight hours after transient transfection, cells showed various morphologic changes including smaller size, nuclear fragmentation, and increased cell debris, suggesting apoptosis in cells overexpressing ARHI (Fig. 2C).

We measured cell proliferation, colony formation, and invasion ability after overexpression of ARHI in PC-3 cells. As shown in Fig. 3A, cell growth was significantly decreased in a time-dependent manner in ARHI overexpressing cells compared with the control cells, suggesting that the gene ARHI has an antiproliferative effect.

We also conducted colony formation assay to assess the long-term effects of ARHI on cell proliferation (Fig. 3B). After 21 days of selection in G418 media, cells transfected with ARHI formed significantly fewer colonies than the control cells, consistent with the cell proliferation assay, further suggesting that ARHI inhibited cell proliferation in prostate cancer cells.

We also examined the effect of ARHI on the invasion ability of prostate cancer cells. As shown in Fig. 3C, cells overexpressing ARHI had 60% of the invasiveness compared with the control cells, indicating ARHI not only affects prostate cancer cell proliferation, but also modulates cell invasion properties. This suggests that downregulation of ARHI may contribute to the metastatic behavior of prostate cancer cells.

Overexpression of ARHI-inhibited prostate cancer cell growth that could be caused by regulation of the cell cycle, increased apoptosis or both mechanisms. To better understand how ARHI effects cancer cell growth, we conducted flow cytometry in PC-3 cells 48 hours after ARHI transfection. As shown in Fig. 4A, cell cycle analysis showed that cells overexpressing ARHI had a significant increase in cells in the G0/G1 phase whereas those in the S phase and G2/M phase decreased. This cell cycle distribution suggests that ARHI causes G0/G1 arrest in cell-cycle progression. To further uncover the molecular mechanism of this kind of growth arrest, we profiled the expression of key cell cycle-related genes in ARHI overexpressing cells and mock controls by using cell cycle PCR arrays from SABioscience. Several genes that were regulated by ARHI had been identified as shown in Fig. 4B. Hect domain and RLD5 (HERC5), Cyclin-dependent kinase inhibitor 1A (CDNK1A), growth arrest and DNA-damage-inducible, alpha (GADD45A) are upregulated more than 5-fold. Protein expression levels of these genes were further confirmed by Western blot (Fig. 4E). Most importantly, CDKN1A and GADD45A function as regulators in the G1 phase of cell cycle progression.

We also examined apoptosis in ARHI overexpressing PC-3 cells by flow cytometric analysis after Annexin-V-FITC-7-AAD staining (Fig. 4C). There was a dramatic decrease in the viable cell population. The apoptotic population went up from 0.03% to 8.16% and early apoptotic cells increased from 0.71% to 1.69% of the cell population after ARHI transfection. We also noticed a large number of dead cells that could be the effects of apoptosis induced by ARHI. The flow cytometric results confirmed the observation of morphologic changes in PC-3 cells after ARHI overexpression showing that ARHI induces apoptosis in prostate cancer cells. Additional studies about gene profiling of apoptosis-related genes were also done. CASP1, harakiri, BCL2 interacting protein (HRK), and tumor necrosis factor superfamily member7 (CD70) were found to be upregulated in both mRNA levels and protein levels (Fig. 4D and 4E). These genes are either involved in apoptosis or play a role in the regulation of apoptosis pathways.

To clarify the mechanism of ARHI downregulation in prostate cancer cells, we investigated the epigenetic regulation of this gene. According to Methprimer software, there are 3 CpG islands in the ARHI promoter and exon region. CpG I is about 300 bp upstream of the transcriptional start site and CpG II and III are downstream of the start site (Supplemental Fig 1A). The methylation status of these ARHI CpG islands was analyzed by using bisulfate DNA sequencing. These data showed ARHI methylation level of approximately 50%–65% in CpG island I and II in immortalized prostate epithelial RWPE-1 cells. However, hypermethylation was observed in CpG island I (about 90%–100%) in prostate cancer cell lines PC-3 (Supplemental Fig. 1B) and Du145 (Supplemental Fig. 1C). PC-3, but not Du145 cells also had hypermethylation in CpG island II (about 65% – 90%). Another typical prostate cancer cell line LNCap did not have increased methylation of the ARHI promoter region (Supplemental Fig. 1D). Furthermore, ARHI expression levels were significantly increased after 5-Aza-2′ dioxycitidine (5-aza; 5μmol/L) treatment for 72 hours in almost all the cell lines except LNCap cells (Supplemental Fig. 1). These findings suggest that frequent hypermethylation of the ARHI promoter region may be a critical event associated with the downregulation of ARHI.

To investigate the new mechanism of ARHI silencing in PC-3 cells, we focused our attention to microRNAs that are a wide class of small, noncoding RNAs that can regulate gene expression by inhibiting translation or inducing cleavage of target mRNAs. We performed a bioinformatic analysis by using miRDB, TargetScan, and PicTar algorithms and predicted 2 microRNAs (miR-221 and miR-222) that might target ARHI in 3′UTR region.

To prove that the ARHI 3′UTR is target of miR-221 and miR-222, we first analyzed their expression in PC-3 and RWPE-1 cells (Fig. 5A). As shown, miR-221 was increased about 20-fold in PC-3 cells compared with RWPE-1 whereas miR-222 was increased 5-fold. These data showed an inverse relationship between ARHI and miR-221 and 222 levels, which was further confirmed by introducing miR-221 and miR-222 inhibitors into PC-3 cells. As shown in Fig. 5B, miR-221 and miR-222 levels were significantly downregulated by the inhibitors, whereas ARHI was significantly increased in both mRNA and protein levels (Fig. 5C). These data strongly support the hypothesis that miR-221 and miR-222 are involved in the regulation of ARHI levels in PC-3 cells.

To demonstrate a direct interaction between miR-221 and miR-222 with ARHI, we cloned a 58 bp (miR-221) and 56 bp (miR-222) ARHI 3′-UTR segment, containing potential target sites for these 2 microRNAs, into the pMIR vector luciferase open reading frame and transfected these reporter construct into PC-3 cells. Empty vectors and vectors contained mutated target sequence were used as control for comparison. Vectors were cotransfected with miR-221 or miR-222 precursor and the β-gal vector that was used for normalization of transfection differences. As shown in Fig. 5D, luciferase activity in PC-3 cells cotransfected with pMIR-ARHI and miR-221 was decreased about 80% compared with the empty vector, whereas cotransfection with pMIR-ARHI and miR-222 decreased luciferase activity over 90%. No significant changes were observed in PC-3 cells transfected with mutant target sequence groups. Taken together, these results provide strong evidence that the ARHI 3′-UTR is a direct target of miR-221 and miR-222 in prostate cancer PC-3 cells.

Genistein, a dietary isoflavone from soy, has been investigated as a potential chemopreventive drug. Here we examined whether genistein could regulate tumor suppressor gene ARHI by altering microRNA levels in prostate cancer cells. We treated PC-3 cells with 50 μmol/L genistein for 4 days and measured miR-221 and miR-222 levels. As shown in Fig. 6A, miR-221 and miR-222 levels decreased about 30% and 55%, respectively after genistein treatment. Meanwhile, ARHI mRNA levels were induced 2-fold after genistein treatment (Fig. 6B). Western blot confirmed that protein levels of ARHI also increased after genistein treatment (Fig. 6B). These data suggest that genistein can partially restore ARHI by decreasing miR-221 and miR-222 levels.

ARHI, a tumor suppressor gene, functions by negative regulation of cell proliferation in several different kinds of tissues (3–6). However, due to its special feature of being a maternally imprinted gene, loss or downregulation of the paternal allele's expression can lead to the unlimited cell growth, which may contribute to tumorigenesis in certain cancers (3–6). Previous studies have showed that ARHI loss or downregulation could be caused by loss of heterozygosity (7) and/or epigenetic events such as DNA methylation, histone methylation, and histone acetylation (9, 10, 15). However, ARHI expression and its function have not been studied in prostate cancer.

In this study, we evaluated ARHI levels in prostate cancer cell lines and prostate cancer tissues. Both RT-PCR and immunohistochemistry showed that ARHI is downregulated in prostate cancer cells and tissues compared with the adjacent normal. And we also anticipated that ARHI staining would be even stronger in normal tissue (nonmalignant tissue) compared with the adjacent normal prostate tissue due to the cancer “field effect” (16). It has been confirmed that multiple genetic and epigenetic changes already occurs in normal adjacent cancers. Difference of ARHI expression between adjacent normal and cancer tissue is consistent with the possibility that this gene might play an important role in regulating cell proliferation in the prostate. We investigated this possibility by overexpression of ARHI in prostate cancer PC-3 cells followed by cell proliferation, invasion, and colony formation assays. Our data showed that in PC-3 cells, ARHI negatively regulates cell growth by causing cell cycle G1/G0 arrest and inducing apoptosis.

We also profiled the expression of key cell cycle and apoptosis pathway genes and identified several components that were unregulated. In the cell cycle pathway, this included upregulation of CDKN1A (also known as P21), GADD45A, and HERC5. The protein that P21 encodes can bind to and inhibit the activity of cyclin-CDK2 or -CDK4 complexes, and thus functions as a regulator of cell cycle progression at G1. Previously, it has also been reported that ARHI overexpression in ovarian cancer is associated with increased expression of p21WAF1/CIP1 (17). GADD45A is a member of a group of genes whose transcript levels are increased following stressful growth arrest conditions and stimulates DNA excision repair, inhibiting cell entry into S phase. HERC5 is a member of the HERC family of ubiquitin ligases and functions as an interferon-induced E3 protein ligase that mediates ISGylation of protein targets. In the apoptosis pathway, CASP1, HRK, and CD70 were upregulated. CASP1 encodes a protease that cleaves IL-1 beta, releasing the mature cytokine that is involved in a variety of inflammatory processes. This gene has been shown to induce cell apoptosis. HRK activates apoptosis and interacts selectively with survival-promoting proteins Bcl-2 and Bcl-X. CD70 binds to CD27 and plays a role in T-cell activation. Taken together, ARHI-mediated regulation of these multiple components of the cell cycle and apoptosis pathways is a novel finding and suggests that ARHI influences cell cycle and apoptosis genes that regulate prostate cancer tumorigenesis.

It is known that aberrant methylation of CpG islands, especially in the promoter regions of genes, serves as an important mechanism for the inactivation of tumor suppressor genes in cancers. Previous reports have shown that ARHI is inactivated by promoter methylation in follicular thyroid carcinoma (5), breast cancers (8), and ovarian cancers (18). Therefore, we studied the ARHI promoter and exon CpG island methylation status in several prostate cell lines. We found that all 3 CpG islands were partially methylated in RWPE-1 cells. In contrast, aberrant methylation of ARHI was observed in prostate cancer cell lines PC-3 and Du145. Especially in CpG island I, hypermethylation (90%–100%) was found in both PC-3 and Du145 cell lines, suggesting methylation in this region could play a critical role in ARHI transcription. Although ARHI expression is downregulated in several different kinds of cancer cell lines, differences were observed between different cell lines in methylation of the CpG island. LNCap cells had a similar methylation status compared with RWPE-1 cells but markedly lower levels of ARHI, indicating methylation is not the only mechanism causing downregulation of ARHI.

We also investigated the mechanism of ARHI downregulation by miRNA regulated pathways. Aberrantly expressed miRNAs are found in many cancer tissues and may be involved in carcinogenesis. Thus, the identification of cancer-specific miRNA targets is critical for understanding the functional role and mechanism. We screened potential miRNAs targeting ARHI with computer software and found more than 10 miRNAs that could potentially target the ARHI gene. Among these miRNAs, miR-221 and miR-222 have been reported to have altered expression in prostate cancer (19, 20). Therefore, we measured their expression levels in PC-3 cells and found increased levels of both miRNAs. Thus these 2 miRNAs represented attractive candidates for our target study. After introducing miR-221 and miR-222 inhibitors into PC-3 cells there was a large increase in ARHI level. To confirm this finding, we performed a luciferase reporter assay with a vector containing the ARHI 3′-UTR target site. The luciferase activity in PC-3 cell transfected with ARHI 3′-UTR was significantly lower than those cells transfected with empty control vector. After transfecting the miR-221 or miR-222 precursor, luciferase activity decreased even more, whereas cells with the control vector showed no measurable difference. Taken together, these results confirmed that miR-221 and miR-222 which are upregulated in prostate cancer cell line can directly target ARHI at the 3′-UTR region, thereby contributing to prostate cancer tumorigenesis.

Genistein, an isoflavone from soy, has been reported to suppress the growth of several prostate cancer cell lines in vitro including PC-3, Du145, and LNCap (21–23) and in TRAMP mice model (24). This suppression has been attributed to regulation of cell cycle and apoptosis pathways (25–29). In this study, we found that genistein regulates ARHI through a novel pathway by regulating miR-221 and 222, thereby affecting ARHI expression levels.

This is the first report to show that ARHI is downregulated in prostate cancer. We also have shown that ARHI is a direct target of miR-221 and 222, which are upregulated in prostate cancer, documenting a new mechanism responsible for ARHI downregulation in prostate cancer. Genistein, a potential nontoxic chemopreventive agent restores expression of ARHI by downregulation of miR-221 and miR-222 and may be an important dietary therapeutic agent for treating prostate cancer.

No potential conflicts of interest were disclosed.

We thank Dr. Roger Erickson for his support and assistance with the preparation of the manuscript.

This study was supported by NIH grants R01CA111470, T32DK007790, and RO1CA138642, and VA Research Enhancement Award Program (REAP) and VA Merit Review grants.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.