Tumor suppressor microRNA miR-145 is commonly down-regulated in colon carcinoma tissues, but its specific role in tumors remains unknown.
Tumor suppressor microRNA miR-145 is commonly down-regulated in colon carcinoma tissues, but its specific role in tumors remains unknown.
In this study, the authors identified the Friend leukemia virus integration 1 gene (FLI1) as a novel target of miR-145. FLI1 is involved in t(11;22)(q24:q12) reciprocal chromosomal translocation in Ewing sarcoma, and its expression appears to be associated with biologically more aggressive tumors.
The authors demonstrated that miR-145 targets a putative microRNA regulatory element in the 3′-untranslated region (UTR) of FLI1, and its abundance is reversely associated with FLI1 expression in colon cancer tissues and cell lines. By using a luciferase/FLI1 3′-UTR reporter system, they found that miR-145 down-regulated the reporter activity, and this down-regulation was reversed by anti–miR-145. Mutation of the miR-145 microRNA regulatory element sequence in the FLI1 3′-UTR abolished the activity of miR-145. miR-145 decreased FLI1 protein but not FLI1 mRNA, suggesting a mechanism of translational regulation. Furthermore, the authors demonstrated that miR-145 inhibited cell proliferation and sensitized LS174T cells to 5-fluorouracil–induced apoptosis.
Taken together, these results suggest that miR-145 functions as a tumor suppressor by down-regulating oncogenic FLI1 in colon cancer. Cancer 2011. © 2010 American Cancer Society.
MicroRNAs (miRNAs) are endogenously expressed noncoding RNAs, 18-25 nucleotides in length, that play important regulatory functions by targeting specific mRNAs1-3 and/or gene promoters.4, 5 They have been implicated as regulatory elements in many cellular processes, including development, heterochromatin formation, and genomic stability in eukaryotes. Most notably, each miRNA may control hundreds of gene targets6, 7 that are involved in disease development. Although the number of verified human miRNAs is still expanding, physiological functions have been determined for only a few of these molecules.
Emerging evidence suggests potential roles of miRNAs as either tumor suppressors or oncogenes.8-12 miR-145, a putative tumor-suppressing miRNA, is down-regulated in a variety of solid tumors, including lung cancer, colorectal cancer, breast cancer, prostate cancer, ovarian cancer, hepatocellular carcinoma, corticotropinoma, and cervical cancer.13-18 Down-regulation of miR-145 has been demonstrated in the lungs of animals exposed to cigarette smoking,19 supporting its involvement in cancer pathogenesis. miR-145 has been associated with inhibition of tumor cell growth both in vitro and in vivo by specifically silencing c-myc or insulin receptor substrate 1 (IRS1).20, 21 In addition, mir-145 has been implicated in the exertion of antineoplastic effects in the lung22 and the inhibition of cell proliferation of a wide range of tumor cells,14, 16, 20, 23 including DLD-1, SW480, HCT116, and LS174T.
Elucidation of the genes targeted by miR-145 has been examined using bioinformatic and proteomic techniques. The IRS1 gene was first evaluated as a potential miR-145 target based on its well-characterized role in tumor biology.21 IRS1, a docking protein for both the type 1 insulin-like growth factor receptor and the insulin receptor, delivers mitogenic, antiapoptotic, and antidifferentiation signals.24-26 As a direct functional mediator of p53, miR-145 also down-regulates proto-oncogene c-myc,20 whose aberrant expression is associated with aggressive and poorly differentiated tumors.27
In this study, we identified the Friend leukemia virus integration 1 gene (FLI1) as a clinically significant target for miR-145. FLI1 has been shown to play an oncogenic role in the promotion of the malignant phenotype. Aberrant expression of FLI1 has been recognized as a seminal event in the initiation of certain types of malignant transformation. The etiology of several virally induced leukemias and of Ewing sarcoma has been associated with FLI1 overexpression.28, 29 By using reporter and functional assays, we demonstrate that miR-145 down-regulates FLI1 by targeting its 3′-untranslated region (UTR), specifically at the miRNA regulatory element of FLI1. By decreasing FLI1 protein levels, miR-145 inhibits cell proliferation in human colon cancer cells and sensitizes LS174T colon cancer cells to 5-fluorouracil–induced apoptosis.
Three human colon carcinoma cell lines were used for this study, including LS174T (Dukes' type B colon cancer), SW620 (Dukes' type C colon cancer), and HCT116 (Dukes' type D colon cancer). All cell lines were obtained from American Type Culture Collection (Manassas, Va). Cells were maintained in Dulbecco modified Eagle medium (Invitrogen, Carlsbad, Calif) supplemented with 10% fetal bovine serum. All cells were incubated at 37°C in a humidified chamber supplemented with 5% CO2. Colon carcinoma specimens and adjacent nontumorous tissues were collected at Renji hospital, Shanghai, China. Patients aged 44 to 75 years were diagnosed as Duke's stage C colon cancers. Fresh tissues were obtained after surgical resection and immediately placed in liquid nitrogen.
All miRNA and ribonucleic acid interference (RNAi) constructs used for cell transfection, including miR-145 precursor, RNAi inhibitor (anti–miR-145), and control RNAi (miR-CT, anti–miR-CT), were purchased from Ambion (Austin, Tex), and sequences were listed at http://www.ambion.com/techlib/resources/miRNA/index.html. Hairpin-it miRNA Real-Time Polymerase Chain Reaction [PCR] Quantification Kits were obtained from GenePharma (Shanghai, China). SYBR Premix Ex Taq (perfect real time) was purchased from Takara Bio (Madison, Wis). FLI1 antibody, Bcl2 antibody, and Rb antibody were purchased from ProteinTech Group (Chicago, Ill). A Cell Counting Kit (CCK-8) was purchased from Dojindo Laboratories (Kumamoto, Japan), and 5-fluorouracil was from Sigma (Milwaukee, Wis).
Transfection of cells was performed with Lipofectamine 2000 Reagent (Invitrogen) following the manufacturer's protocol. Briefly, the cells were seeded in 6-well plates at 30% confluence the day before transfection. miR-145 precursor, miR-145 inhibitor, and miRNA control (50 nM each), were used for each transfection.
We collected tumor and adjacent nontumor samples from colon cancer patients. Total miRNA was extracted with an mirVana miRNA Isolation Kit (Ambion) using the miRNA enrichment protocol. Mature miR-145 was measured by Northern blotting using a NorthernMax-Gly Kit (Ambion), following the protocol provided by the manufacturer. Briefly, after RNA electrophoresis, the transferred membrane was prehybridized with ULTRAhyb, and detected with a miR-145–specific oligonucleotide probe labeled with digoxigenin-ddUTP using a DIG Oligonucleotide 3′-End Labeling Kit (Roche Diagnostics, Indianapolis, Ind).
Hairpin-it miRNAs Real-Time PCR Quantification Kit uses a stem loop approach to detect expression of mature miR-145. For the reverse transcript reaction, 200 ng of total RNA was used and mixed with the miR-145 reverse transcriptase primer. The reaction was carried out for 30 minutes at 16°C, 30 minutes at 42°C, and 5 minutes at 85°C, and then held at 4°C. For the PCR reaction, 2 μL cDNA products were used along with an miR-145–specific primer set. The PCR reaction was conducted at 94°C for 3 minutes, followed by 45 cycles of 94°C for 20 seconds, 50°C for 25 seconds, and 72°C for 20 seconds in a Rotogene 3000 real-time PCR system (Corbett Research, Sydney, Australia). U6 RNA was used for normalization.
To detect relative levels of FLI1 mRNA, real-time PCR was performed using the SYBR method at the following conditions: 95°C for 30 seconds, 1 cycle; 95°C for 5 seconds, 60°C for 20 seconds, and 72°C for 15 seconds, 40 cycles. PCR primers were FLI1 sense, 5′-CAG TCG CCT AGC CAA CCC TG and antisense, 5′-GCA ATG CCG TGG AAG TCA AAT.
Putative miR-145 targets were predicted by the Target Scan program (http://www.targetscan.org), the miRanda program (http://microrna.sanger.ac.uk), and the PicTar (4-way) program (http://pictar.bio.nyu.edu).30-32 The data obtained from these 3 programs were integrated with the miRGen Target program (http://www.diana.pcbi.upenn.edu/miRGen.html).33
After bioinformatic analysis, the putative miRNA regulatory element (miRNA regulatory elements, Table 1) of miR-145 was chemically synthesized and cloned into pGL3 control vector (Promega, Madison, Wis) at the Xba1 site. To construct the Luc-FLI1-3′UTR-full vector, full-length 3′UTR of FLI1 was amplified from the cDNA of LS174T cells using FLI1 PCR primers—sense 5′ CTA GAG AAG CCC ATC CTG CAC ACT 3′ and antisense 5′ CTA GAC GTT GTT TTT CCC AGA GCT 3′—and then cloned into the pGL3 control vector at the Xba1 site. To create the mutated FLI1-3′UTR vector, 8 nucleotides (GUG AAG UUU UCA CCC ggt gat cg) were changed for the reporter construct.
293T cells were seeded in 24-well plates and transfected with luciferase reporters containing the putative miRNA regulatory element for miR-145, miR-145 precursor, and miR-145 inhibitor. Transfection efficiency was corrected by a Renilla luciferase vector (pRL-CMV, Promega). The cells were harvested for luciferase assays 24 hour after transfection. A luciferase assay kit (Promega) was used to measure the reporter activity according the manufacturer's protocol.
Total protein was isolated from LS174T cells transfected with miR-145 precursor or inhibitor. Protein concentration was measured using Pierce BCA Protein Assay Reagent (Thermo-Fisher Scientific, Rockford, Ill). Cell lysates (50 μg) were electrophoresed through 10% polyacrylamide gels and transferred to a nitrocellulose membrane. The membrane was incubated with FLI1 antibody, Bcl2 antibody, or Rb antibody. Secondary antibodies were labeled with IRDyes. Signals were observed using an Odyssey Infrared Imaging System.
Measurement of viable cell mass was performed with a cell counting kit that counts living cells using WST-8. According to the protocol, 100 μL of LS174T cells treated with 50 nM miR-145 precursor or 100 nM small interfering RNA (siRNA)/FLI1 were plated on 96-well plates. Ten microliters of Cell Counting Assay Kit-8 solution was added to each well, and the absorbance was measured at 450 nm using a microplate reader. The amount of the formazan dye, generated by the activities of dehydrogenases in cells, is directly proportional to the number of living cells.
LS174T cells were transfected with 50 nM miR-145 precursor or 100 nM siRNA/FLI1 for 24 hours, and then treated with 10 μmol/L 5-fluorouracil for 24 hours. Cells were harvested and resuspended with 500 μL of binding buffer. The cell suspension (100 μL) was incubated with 5 μL annexin-V and propidium iodide at room temperature for 20 minutes. The stained cells were analyzed with fluorescent-activated cell sorting (FACS) using BD LSR II flow cytometry (BD Biosciences, San Diego, Calif).
To study the role of miR-145 in tumors, we initiated a screening of its putative gene targets using 3 different computational methods: the Target Scan program, Miranda program, and PicTar (4-way) program. By using these bioinformatics prediction models, we initially obtained >977 miR-145 targets from the Miranda target search program, >396 targets from the Target Scan program, and >326 targets from the PicTar (4-way) program (data not shown), which were far more than the average prediction of about 100 target genes per single miRNA.7 Only a small fraction of predicted targets are true miR-145 targets. Thus, we attempted to improve the prediction accuracy by integrating the data acquired from all 3 of the abovementioned programs using the miRGen Target program, obtaining roughly 47 targets for miR-145 (data not shown). Among them, 5 genes containing 7 putative binding sites (miRNA regulatory elements) (Table 1) in the gene structure were chosen for further study, including C11orf9, FLI1, CPEB4, FZD7, and CBFB.
Among these 5 selected genes, C11orf9 belongs to a previously uncharacterized family of transcription factors and is strongly induced in invasive or metastatic tumor cells.34FLI1 is overexpressed in virally induced leukemias and in Ewing sarcoma as a result of a characteristic t(11;22)(q24:q12) reciprocal chromosomal translocation.35, 36 CPEB is a sequence-specific RNA-binding protein that mediates many processes including germ-cell development, cell division, and cellular senescence.37 Up-regulated FZD7 leads to activation of the Wnt/beta-catenin/TCF pathway in human gastric cancer.38 Similarly, the overexpressed transcription factor CBFB is an independent predictor of tumor survival in colorectal cancer patients.39
We chemically synthesized these 7 miRNA regulatory elements and tested their functions by cloning them into the Xba1 site of pGL3 reporter vector (Fig. 1A), downstream of the reporter gene luciferase. To test the response of each miRNA regulatory element to miR-145, luciferase/miRNA regulatory element constructs were cotransfected with miR-145 into 293T cells. By using this reporter system, miR-145–responsive miRNA regulatory elements were identified by simply measuring luciferase activity. Two miRNA regulatory elements were located in FLI1 mRNA: an upstream (FLI1-94) site and a downstream (FLI1-519) site. As shown in Figure 1B, only the downstream FLI1 miRNA regulatory element (FLI1-519) responded to miR-145. Notably, luciferase reporter activity decreased >50% when the FLI1 miRNA regulatory element was inserted at the 3′UTR of the reporter gene in the Luc-FLI1-519 construct. The upstream FLI1 miRNA regulatory element (FLI1-94), however, did not respond to miR-145. FLI1-519, referred to as FLI1 miRNA regulatory element, was thus used for further study.
By using a luciferase reporter system, we then assessed whether the putative binding sites for miR-145 altered the activity of the upstream luciferase reporter gene. With the luciferase/FLI1 miRNA regulatory element system, we found that the inhibition of luciferase activity by miR-145 was dose dependent (Fig. 2A). To test the specificity of miR-145, we cotransfected cells with an RNAi inhibitor (anti–miR-145) that is an authentic blocker of miR-145. We showed that anti–miR-145 specifically abolished the miR-145 induced-inhibition of luciferase activity (Fig. 2B).
To further demonstrate the importance of the FLI1 miRNA regulatory element, a substitution mutation was generated to test its activity. In the FLI1-3′UTR-M vector, 8 nucleotides (AACUGGAA) were replaced with GGTGATCG (Fig. 2C). We cloned the full-length of the FLI1 3′-UTR downstream of the reporter. As expected, down-regulation of reporter activity was detected in the construct that contains the full length of the FLI1 3′-UTR. Correspondingly, we demonstrated that the mutation in the miRNA regulatory element abolished the miR-145–mediated inhibition of the reporter gene (Fig. 2D). Taken together, these data suggest that the miR-145 binding site present in the FLI1 3′UTR is critical for miR-145–mediated gene regulation.
To seek the link between miR-145 and FLI1, we measured the endogenous expression of mirR-145 and FLI1 in colon cancer tissues and LS174T cells. We used Northern blotting to examine the expression of miR-145 in paired colon tumors and adjacent normal tissues collected from the same patients. Expression of miR-145 was also examined in 3 human colon cancer cell lines, LS174T, SW620, and HCT116. As seen in typical samples in Figure 3A, colon cancer tissues showed a significant decrease in mature miR-145 abundance in comparison with the paired adjacent nontumor tissues. Three colon cancer cell lines also showed down-regulation of miR-145, especially in LS174T cells. The low abundance of miR-145 in colon tumor tissues is consistent with a potential suppressor role of the miRNA in cancer development.
In contrast, FLI1 was overexpressed in the same colon tumor tissues and LS174T cells (Fig. 3B) where miR-145 was down-regulated. The inverse correlation suggests that miR-145 and FLI1 are coordinately regulated in the development of tumors.
Translational repression is a major mechanism of miRNAs to regulate gene expression in mammals.40 To determine whether miR-145 uses the same mechanism, LS174T cells that express very little miR-145 but overexpress FLI1 were transfected with miR-145 or its inhibitor. The ectopic expression of miR-145 was confirmed by Hairpin-it miRNAs Real-Time PCR Quantification Assay. As expected, about 3-fold increase in mature miR-145 was detected in the miR-145 precursor-transfected cells (Fig. 4A). In contrast, transfection with anti–miR-145 reduced miR-145 by almost 50% in LS174T cells (Fig. 4B).
We then measured the miR-145 target gene FLI1 in the transfected LS174T cells. Ectopic expression of miR-145 significantly reduced FLI1 protein at 48 hours (Fig. 4C, lane 2). This inhibition was abolished in naive cells by the antagonism of miR-145 using miR-145 inhibitor (Fig. 4C, lane 4). However, we did not detect the inhibition of FLI1 at the mRNA level as measured by real-time PCR (Fig. 4D). Similarly, neither was FLI1 mRNA affected by miR-145 inhibitor (Fig. 4E). These results suggest that miR-145 targets FLI1 by functioning on translational regulation.
We were then interested in whether this down-regulation of FLI1 by miR-145 was able to affect tumor cell growth. miR-145 and FLI1 siRNA were transfected in LS174T cells that expressed low miR-145 but high FLI1. The cell growth was determined at different time points after transfection by WST-8 assays. As shown in Figure 5A, transfection with miR-145 inhibited cell proliferation as compared with control cells. Similarly, blockage of the miR-145 target gene FLI1 using siRNA also deceased tumor cell growth. In general, tumor cells transfected with miR-145 grew much more slowly than those with control miRNA in parallel with the down-regulation of FLI1 proteins (Fig. 5B).
As an initial step to delineate the role of miR-145, we also measured additional 2 genes, Rb and Bcl-2, which are commonly involved in the regulation of cell growth. We found that decreased expression of FLI1 by miR-145 and FLI1 siRNA was also associated with up-regulated Rb and down-regulated Bcl-2 (Fig. 5B). Altered expression of Rb and Bcl-2 protein may contribute to the inhibition of LS174T cell proliferation.
We speculated that down-regulation of the antiapoptotic Bcl-2 would enhance apoptosis and would sensitize tumor cells to colon cancer chemotherapy drugs, such as 5-fluorouracil. We thus assessed apoptosis by FACS after the cotreatment of miR-145 and 5-fluorouracil. As shown in Figure 5C, inhibition of FLI1 by either miR-145 or FLI1 siRNA sensitized cells to 5-fluorouracil–induced apoptosis. We detected an approximately 2-fold increase in apoptosis induction by miR-145 as compared with 5-fluorouracil alone. These data further support miR-145 as a tumor suppressor by indicating that it participates in apoptosis regulation.
Aberrant expression of miR-145 tumor suppressor has been implicated in a variety of cellular pathways involved in carcinogenesis. However, the exact role of miR-145 in tumorigenesis is still unclear, largely because of limited knowledge about miR-145 targets. In this study, we used bioinformatics prediction models to identify FLI1 as a novel target of miR-145 in tumors. In our colon cancer samples, the increased expression of the FLI1 gene in colon tumors was in parallel with decreased miR-145. By using a luciferase reporter system, we showed that the miR-145 binding site present in the FLI1 3′UTR region is critical for miR-145–mediated regulation.
FLI1 has been well documented to act as an oncogene in tumor development, especially in Ewing sarcoma. The etiology of several virally induced leukemias as well as human Ewing sarcoma has been associated with FLI1 overexpression.41, 42 The rate-limiting oncogenic mutation in Ewing sarcoma has been identified as a chromosomal translocation, t(11;22)(q24;q12), that leads to the expression of a chimeric transcription factor, EWS-Fli1.28, 29 Ectopic expression of FLI1 has also been shown to be associated with decreased expression of Rb, decreased expression of GATA1, and increased expression of Bcl-2, possibly leading to a block in apoptosis and differentiation.35, 41-44 Aberrant expression of FLI1 has been recognized as a seminal event in the initiation of certain types of malignant transformation. Indeed, the etiology of several virally induced leukemias, including Friend virus-induced erythroleukemia, has been associated with FLI1 overexpression.41 The clinical relevance of FLI1 becomes apparent in human Ewing sarcoma, in which FLI1 is the target of a characteristic chromosomal translocation.42FLI1 is expressed at high levels in Friend murine leukemia virus–induced erythroleukemias,45 malignant melanoma, small cell carcinomas of the lung, and adenocarcinomas.46, 47
However, the role of FLI1 in colon cancer, especially its link with miR-145, has not been explored. We wished to examine how FLI1 was regulated by miR-145. Ectopic expression of miR-145 significantly reduced FLI1 protein at 48 hours in LS174T cells (Fig. 4C), and this inhibition was equally efficient to those by using Fl1-1 siRNA (Fig. 5B). However, we did not observe any effects on the FLI1 mRNA level (Fig. 4D, E). Although miRNAs may regulate protein expression by accelerating messenger RNA degradation,2 miR-145 is presumably acting to block translation of FLI1, which is believed to be the most common mechanism of miRNA targeting.48
It was also important to note that miR-145 was more effective than FLI1 siRNA in inhibiting LS174T cell proliferation (Fig. 5A), despite the finding that the down-regulation of FLI1 protein levels by miR-145 was not more effective than siRNA (Fig. 5B). In addition, miR-145 and siRNA have similar effects on sensitizing the LS174T cells to 5-fluorouracil–induced apoptosis (Fig. 5C). These findings suggest that other miR-145 targets may also play an important role in inhibiting cellular proliferation. In addition, as shown in Figure 5B, decreased expression of FLI1 by miR-145 is associated with increased expression of Rb and decreased expression of Bcl-2. The altered expression of Rb and Bcl-2 protein may in part account for the role of miR-145 in inhibiting tumor cell growth.
In summary, this study identified FLI1 as a novel gene target of miR-145 in colon cancers. These results demonstrate that miR-145 acts as a tumor suppressor, targeting the 3′-UTR of FLI1 mRNA. Down-regulation of FLI1 has a profound effect on the growth of human colon cancer cells. Its inhibition of growth in colon cancer cells is compatible with the well-known ability of FLI1 to stimulate cell proliferation and transformation. miR-145 regulates FLI1 at the protein expression level, although the specific mechanism requires further investigation. These studies extend our knowledge concerning FLI1 as an oncogene involved in colon tumorigenesis, although the specific role needs further investigation. Given that a single miRNA may have multiple targets, we believe that miR-145 also has many other unknown targets. It is our expectation that more miR-145 targets will be identified in the near future, advancing our understanding of the molecular basis of miR-145–mediated tumorigenesis. In addition, miR-145 as a potential therapeutic molecular agent in colon cancer and other tumors, like Ewing sarcoma, may merit further investigation.
Supported by The National Key Program for Basic Research of China (2010CB529902 and 2010CB834201), The National Natural Science Foundation of China (30973663 and 10935009 to G.Q.); The Science and Technology Commission of Shanghai (08ZR1412500 and 10JC1409100) , The Shanghai Leading Academic Discipline Project (S30205), The Innovation Program of Shanghai Municipal Education Commission (jdy08054,09ZZ110 to S.G.); and NIH grant (1R43 CA103553-01), and The Department of Defense Grant (W81XWH-04-1-0597 to J.F.H.).