microRNAs (miRNA) are small noncoding RNAs that participate in diverse biological processes by suppressing target gene expression. Altered expression of miR-21 has been reported in cancer. To gain insights into its potential role in tumorigenesis, we generated miR-21 knockout colon cancer cells through gene targeting. Unbiased microarray analysis combined with bioinformatics identified cell cycle regulator Cdc25A as a miR-21 target. miR-21 suppressed Cdc25A expression through a defined sequence in its 3′-untranslated region. We found that miR-21 is induced by serum starvation and DNA damage, negatively regulates G1-S transition, and participates in DNA damage–induced G2-M checkpoint through down-regulation of Cdc25A. In contrast, miR-21 deficiency did not affect apoptosis induced by a variety of commonly used anticancer agents or cell proliferation under normal cell culture conditions. Furthermore, miR-21 was found to be underexpressed in a subset of Cdc25A-overexpressing colon cancers. Our data show a role of miR-21 in modulating cell cycle progression following stress, providing a novel mechanism of Cdc25A regulation and a potential explanation of miR-21 in tumorigenesis. [Cancer Res 2009;69(20):8157–65]
- cell cycle
- DNA damage
- colon cancer
microRNAs (miRNA) are evolutionarily conserved, 20- to 25-nucleotide-long, noncoding RNAs that bind to their targets through partial complementary sequence recognition. This results in either degradation of mRNA or inhibition of translation, thus modulating expression of miRNA targets ( 1). Several hundred miRNAs have been identified in human cells ( 2). It is estimated that a single miRNA can regulate hundreds of targets, and ≥30% of human mRNAs are regulated by miRNAs ( 1, 2). Therefore, it is not surprising that miRNAs are involved in diverse biological processes, including cell differentiation, proliferation, and apoptosis, presumably through a myriad of targets ( 2).
Deregulation of miRNAs contributes to human pathogenesis including cancer ( 2). For example, aberrant expression of miRNAs, including miR-21, miR-17-92, miR-15, miR-16, and let-7, has been reported in cancer ( 3). Furthermore, a substantial number of miRNA genes are located in the fragile sites in the genomic regions that are frequently amplified, deleted, or rearranged in cancer, providing plausible mechanisms of deregulated expression ( 4, 5). A theme is emerging that a miRNA can be considered either a tumor suppressor or an oncogene depending on its targets in different tissues and cell types ( 6– 8). Identification of relevant targets or pathways controlled by miRNAs will ultimately provide insights into their biological functions.
Altered expression of miR-21 has been reported in cancer. For example, miR-21 was reported to have substantially higher expression in normal tissues than in colon cancers or in NCI-60 tumor cell lines ( 8). On the other hand, miR-21 is overexpressed in cancers of the breast, lung, pancreas, prostate, stomach, and brain ( 9, 10). Higher expression of miR-21 was found in colon adenocarcinomas than in the normal mucosa and was associated with decreased overall survival ( 11). A limited number of genes, including PTEN, TPM1, Pdcd4, Spry1, and Spry2, have been reported to be targets of miR-21, suggesting potential functions in regulating cell proliferation, apoptosis, and invasion ( 12– 16). However, the precise role of miR-21 in cancer remains to be defined.
The cell division cycle 25 (Cdc25) family of proteins are highly conserved dual-specificity phosphatases that dephosphorylate and activate cyclin-dependent kinase complexes. Three isoforms have been identified in mammalian cells, Cdc25A, Cdc25B, and Cdc25C ( 17). Overexpression of Cdc25 family proteins, mostly Cdc25A and Cdc25B, correlates with more aggressive disease and poor prognosis in some cancers and leads to genetic instability in mice ( 18, 19). Cdc25A positively regulates G1-S and G2-M transitions by activating distinct cyclin/cyclin-dependent kinase complexes ( 18, 19). Moreover, timely inactivation of Cdc25A facilitates checkpoint activation on DNA damage. Cdc25A activities are tightly regulated by multiple mechanisms during the cell cycle, and ubiquitin-mediated proteolysis is the major mechanism of Cdc25A turnover ( 17). For example, hyperphosphorylation of Cdc25A by the ATR-Chk1 signaling leads to its degradation and contributes to a delay in the cell cycle, which allows either DNA repair or apoptosis, depending on the extent of DNA damage ( 17, 19, 20).
In the current study, we reported a novel role of miR-21 in modulating cell cycle progression and DNA damage checkpoint activation via Cdc25A. Cdc25A was identified and validated as a miR-21 target using miR-21 knockout colon cancer cell lines. miR-21 was found to be induced by serum starvation, negatively regulates G1-S transition, and participates in DNA damage checkpoint activation in response to γ-irradiation. Our data provide a novel mechanism of Cdc25A mRNA turnover and a potential role of miR-21 deregulation in tumorigenesis.
Materials and Methods
Targeting the miR-21 gene. Gene targeting vectors were constructed using a recombinant adeno-associated virus (AAV) system as described ( 21– 23) with minor modifications. Briefly, two homologous arms flanking the miR-21 locus, which are 1.17 and 1.15 kb, respectively, along with the neomycin-resistant gene cassette (Neo), were inserted between two NotI sites in the AAV shuttle vector pAAV-MCS (Stratagene) by a four-way ligation reaction. Packaging of recombinant AAV was done by using the AAV Helper-Free System (Stratagene) according to the manufacturer's instructions. RKO and DLD1 cells were infected with recombinant AAV and selected by G418 (0.4 mg/mL) for 3 weeks. G418-resistant clones were screened by PCR for targeting events with primer pairs listed in Supplementary Table S1 using pooled genomic DNA ( 24). The same targeting construct was used in the second round of gene targeting following the excision of Neo gene flanked by LoxP sites in a heterozygous clone with an adenovirus expressing Cre recombinase ( 24). After the second round of gene targeting, Neo was excised by adenovirus expressing Cre recombinase infection again, and gene targeting was verified by genomic PCR, reverse transcription-PCR (RT-PCR), and Northern blot. The detailed procedures of gene targeting and PCR screening are available upon request, and the primers used are listed in Supplementary Table S1.
Microarray analysis. Total RNA was isolated 48 h following transfection from cells cultured in T25 flasks. Microarray analysis was done and relative gene expression was analyzed as described previously by the Core Facility at the University of Pittsburgh School of Medicine ( 25).
Cell culture. Human colorectal cancer cell lines RKO and DLD1 were obtained from the American Type Culture Collection and cultured in McCoy's 5A modified medium (Invitrogen) supplemented with 10% defined fetal bovine serum (Hyclone) and 100 units/mL penicillin and 1% streptomycin (Invitrogen). Cells were maintained at 37°C with 5% CO2. In some experiments, cells were grown in medium containing 0.5% serum. Details on serum-stimulated G1-S transition, radiation-induced transient G2-M checkpoint, and clonogenic survival are described in Supplementary Material.
Isolation of miRNAs, real-time PCR assays, and Northern blotting for mature miRNAs. The expression of mature miRNAs was determined by real-time PCR ( 26) and Northern blot. The expression of protein coding mRNAs was quantitated by real-time PCR. Details are described in Supplementary Material.
Transfections. Transfection with 100 nmol/L pre-miR-21, 100 nmol/L anti-miR-21 (Ambion), or 200 nmol/L Cdc25A small interfering RNA (siRNA; Dharmacon) was done with Lipofectamine 2000 (Invitrogen) according to the manufacturers' instructions. The target sequence for Cdc25A is GGAAAAUGAAGCCUUUGAG ( 27). Cells plated at 20% to 30% confluence in 6-well plates were transfected twice in 48 h and split into T25 flask 10 h after the second transfection. The next day, the cells in T25 were either subjected to serum starvation and stimulation or irradiated as described above.
Luciferase reporter constructs. The reporter constructs containing the 3′-untranslated region (3′-UTR) of Cdc25A were cloned into the pMIR-REPORT vector (Ambion) using PCR-generated fragment. Site-directed mutagenesis with the QuikChange XL Site-Directed Mutagenesis Kit (Stratagene) was used to introduce mutations in the miR-21 binding site. Reporter assays were carried out as described previously with a transfection control ( 28). All the experiments were done in triplicate and repeated at least three times on different days. Details were described in Supplementary Material (Supplementary Table S2).
Bromodeoxyuridine incorporation and mitotic index. Bromodeoxyuridine (BrdUrd) incorporation was analyzed by microscopy or flow cytometry following staining with anti-BrdUrd, Alexa Fluor 488–conjugated antibody. Mitotic index was measured by phosphorylated histone 3 staining. Detailed methods are described in Supplementary Material.
Western blotting. Western blotting was done as described previously ( 29). The antibodies used for Western blotting included those against Cdc25A, Cdc25C, Cdc2, cyclin B1, Chk1 (Santa Cruz Biotechnology), phosphorylated histone H3 (Millipore), α-tubulin (EMD Biosciences), phosphorylated Cdc2 (Cell Signaling), Cdc25B (BD Biosciences), and β-Trcp (Invitrogen). Quantification of relative expression was determined by densitometry as described ( 30).
Statistical analysis. Statistical analysis was done using GraphPad Prism IV software. P values were calculated by Student's t test. P values <0.05 were considered significant. Mean ± SD is displayed in the figures.
Targeted deletion of miR-21 in colon cancer cells. Aberrant expression of miR-21 has been reported in colon cancer ( 8). We were interested in determining its potential role in tumorigenesis by identifying miR-21 targets. To avoid the limitations of down-regulating miRNA expression with antisense oligos ( 31), we knocked out the miR-21 precursor sequence in RKO and DLD1 colorectal cancer cells using the recombinant AAV system ( Fig. 1A ; refs. 21, 22). Both of these lines express relatively high levels of miR-21 ( 8, 32). After two rounds of homologous recombination, miR-21 knockout clones were identified by PCR amplification of the corresponding genomic regions ( Fig. 1B; Supplementary Fig. S1A). RT-PCR and Northern blot confirmed that the mature miR-21 was not expressed in these knockout clones ( Fig. 1C and D).
Identification of potential miR-21 targets through microarray analysis. miRNAs regulate their target genes via mRNA degradation and/or inhibition of translation ( 33). Their potential targets can be identified using high-throughput methods such as microarray analysis ( 34). To identify potential miR-21 targets predicted to have elevated expression in miR-21 knockout cells, we performed microarray analysis on RKO parental cells transfected with pre-miR-21 and miR-21 knockout cells transfected with a control siRNA for 48 h. Over 100 candidates showed at least a 2-fold increase in their expression in miR-21 knockout cells (Supplementary Table S3; P ≤ 0.02). Interestingly, several proteins involved in cell cycle and DNA damage responses were among them. We then chose 12 candidates and validated 7 of them by quantitative RT-PCR analysis (Supplementary Fig. S2). Cdc25A was chosen for further analysis due to its well-established role in cell cycle regulation and cancer.
Cdc25A expression is regulated by miR-21 in colon cancer cells. To validate Cdc25A as a miR-21 target, its mRNA and protein levels were compared in parental and miR-21 knockout RKO cells. Consistent with the results obtained in microarray analysis, Cdc25A levels were significantly up-regulated in miR-21 knockout cells ( Fig. 2A and B ). Transfection of pre-miR-21 decreased the levels of Cdc25A transcript and protein ( Fig. 2C and D; Supplementary Figs. S1B and S3). These findings were confirmed in DLD1 parental and miR-21 knockout cells (Supplementary Fig. S3A-D). miR-21 deletion did not affect the expression of the other two Cdc25 family members, Cdc25B and Cdc25C, or its established regulators Chk1 or β-TrCP (Supplementary Fig. S4A). Transient transfection of anti-miR-21 also elevated Cdc25A expression in RKO cells as did miR-21 targeting ( Fig. 2D, right). To examine the expression of miR-21 in relation to Cdc25A in cancer, we analyzed the expression of miR-21 in 12 colon cancers that overexpress Cdc25A using matched normal and tumor tissues ( 35). miR-21 was found to be underexpressed in 6 of 12 (50%) tumors (range, 2- to 7-fold; Fig. 3 ).
Cdc25A is a miR-21 target. On a closer inspection, a putative miR-21 binding site located in the 3′-UTR of Cdc25A gene was predicted by two algorithms (TargetScan and PicTar; Fig. 4A ). Importantly, this putative miR-21 binding site is 100% conserved in five species in the region that pairs with the seed sequence ( Fig. 4A). The 3′-UTR of Cdc25A containing this site was cloned into pMIR-REPORT miRNA reporter vector. The luciferase activities of this reporter in miR-21 knockout cells were ∼60% higher than that in parental RKO cells but were suppressed by pre-miR-21 transfection ( Fig. 4B), suggesting a regulatory element in its 3′-UTR. We then mutated the miR-21 binding site in the reporter construct Luc-Cdc25A-Mut-UTR and found its activities were similar in parental and miR-21 knockout RKO cells ( Fig. 4). Transfection of pre-miR-21 did not decrease the activities of the mutant reporter in either parental or miR-21 knockout cells ( Fig. 4B), suggesting specificity of this sequence. We also examined the expression of several reported miR-21 targets in the microarray data or by Western blotting, including PTEN, Pdcd4, Bcl-2, TMP1, Spry1, and Spry2 (Supplementary Table S4; Supplementary Fig. S4B). Only Spry1 and Spry2 appear to be significantly up-regulated (1.8- and 1.44-fold) in miR-21 knockout cells but not the other three genes (Supplementary Table S4; Supplementary Fig. S4B). Together, these results indicate that miR-21 regulates Cdc25A through the miR-21 binding site in its 3′-UTR and establish Cdc25A as a direct target of miR-21.
miR-21 inhibits cell proliferation following serum starvation and delays G1-S transition through Cdc25A. Cdc25A is an important regulator of cell cycle progression during G1-S transition ( 36, 37). To evaluate whether miR-21 affects cell cycle progression, we compared the growth rate of parental and miR-21 knockout cells under normal serum (10%) and low-serum condition (0.5%) over a course of 7 days. The growth rate of parental and miR-21 knockout cells was indistinguishable under the normal serum condition in the entire 7 days (Supplementary Fig. S4C). However, RKO miR-21 knockout cells exhibited enhanced proliferation over wild-type cells in the low-serum condition ( Fig. 5A, top ). Under these conditions, no significant levels of apoptosis were detected in either wild-type or miR-21 knockout cells (data not shown). Using quantitative RT-PCR, we found that miR-21 levels were induced 2- to 10-fold by serum starvation in wild-type cells starting at 24 h ( Fig. 5A, middle). Serum starvation also caused an apparent reduction in Cdc25A levels, which was significantly blunted in miR-21 knockout cells ( Fig. 5A, bottom). The induction of miR-21 in parental cells and elevated Cdc25A and enhanced proliferation in miR-21 knockout cells were also observed following complete serum starvation (0% serum; data not shown).
Because serum has been well documented to stimulate the G1-S transition, we therefore specifically evaluated a potential role of miR-21 in this process ( 17, 27, 38). Parental and miR-21 knockout RKO cells were serum starved for 48 h and subsequently stimulated with 10% FCS. The cell cycle profiles were followed by flow cytometry in a time-course experiment. Serum addition induced a higher degree of S-phase entry in miR-21 knockout cells compared with parental RKO cells. The effects were most pronounced at 15 h (17% versus 7%) and 16 h (16% versus 10%) and gradually diminished ( Fig. 5B). BrdUrd staining indicated increased DNA synthesis in miR-21 knockout cells compared with parental cells ( Fig. 5C; Supplementary Fig. S5A). Consistent with an accelerated entry into S phase, high levels of Cdc25A were detected in miR-21 knockout cells as early as 15 h (Supplementary Fig. S5B). Transfection of pre-miR-21 significantly reduced DNA synthesis and Cdc25A levels in miR-21 knockout cells ( Fig. 5D). No significant difference in the cell cycle distribution was found between wild-type and miR-21 knockout RKO or DLD1 cells growing in log phase with 10% serum (Supplementary Fig. S5C). Taken together, these results indicate that miR-21 induction inhibits the G1-S transition by suppressing Cdc25A expression.
miR-21 modulates DNA damage–induced G2-M checkpoint through Cdc25A. Cdc25A has been shown to regulate the G2-M transition and its inactivation is critically involved in establishing a G2-M checkpoint following γ-irradiation ( 17, 27, 38). We tested whether miR-21 is involved in establishing a transient G2-M checkpoint 1 h following irradiation by analyzing phosphorylated histone H3–positive cells ( 39). As expected, radiation (12 Gy) induced a >90% drop in mitotic cells, which was significantly inhibited in miR-21 knockout RKO cells ( Fig. 6A ; Supplementary Fig. S6A). Using quantitative RT-PCR, we found that miR-21 levels were induced in parental RKO cells within 1 h of γ-irradiation ( Fig. 6A, right). miR-21 knockout DLD1 cells also exhibited a defective G2-M checkpoint associated with elevated levels of Cdc25A (Supplementary Fig. S6B and C). Transfection of pre-miR-21 or Cdc25A siRNA significantly suppressed the fraction of mitotic cells in both wild-type and miR-21 knockout cells, but the inhibition was more pronounced in miR-21 knockout cells ( Fig. 6B). Transfection of anti-miR-21 only elevated the fraction of mitotic cells in wild-type cells but did not further elevate that in miR-21 knockout cells ( Fig. 6B). Cdc25A levels rapidly decreased following radiation in both parental and knockout cells but were substantially higher in miR-21 knockout cells ( Fig. 6C). As expected, transfection of pre-miR-21 or Cdc25A siRNA decreased the levels of Cdc25A, whereas anti-miR-21 elevated those in wild-type cells ( Fig. 6C).
To determine whether miR-21 affects radiosensitivity, we evaluated the clonogenic survival of parental and miR-21 knockout RKO cells. miR-21 knockout RKO cells were found to have increased clonogenic survival following several doses of irradiation ( Fig. 6D). miR-21 has been reported previously to regulate cell proliferation and apoptosis in glioblastoma and breast cancer cells ( 9, 13, 40). However, miR-21 did not appear to affect apoptosis induced by a variety of anticancer agents in either RKO or DLD1 cells (Supplementary Fig. S4D; data not shown). The regulation of Cdc25A by miR-21 appears to be independent of the tumor suppressor p53, as it occurs in both p53 wild-type RKO cells and p53 mutant DLD1 cells. The above results suggest that miR-2-mediated down-regulation of Cdc25A contributes to the activation of the G2-M checkpoint following radiation.
Our study provides a novel function of miR-21 in regulating cell cycle progression and checkpoint activation through Cdc25A in colon cancer cells. This conclusion is supported by several lines of evidence: increased expression of Cdc25A in miR-21 knockout RKO and DLD1 cells, which is suppressed by expression of pre-miR-21; a putative miR-21 binding site in the 3′-UTR that is subject to miR-21 regulation; the induction of miR-21 by serum starvation and DNA damage, accelerated G1-S transition in miR-21 knockout cells; and compromised G2-M checkpoint in response to γ-irradiation, all of which were partially rescued by pre-miR-21 or Cdc25A knockdown.
The major mechanism of rapid turnover of Cdc25 family proteins is regulated by ubiquitin-mediated proteolysis ( 19). Our findings suggest that the full extent of Cdc25A inactivation requires miR-21 in colon cancer cells, which represents a novel mechanism of Cdc25A mRNA turnover. An involvement of miR-21 in cell cycle progression following stress is supported by several recent studies, as it was induced by the chemotherapeutic drug 5-fluorouracil in colon cancer cells ( 41) and by UV irradiation in primary fibroblasts ( 42) or in colon cancer cells (Supplementary Fig. S6D). Cdc25A contains a large number of phosphorylation sites recognized by cyclin-dependent kinase 1, Chk1/Chk2, and p38 ( 17, 19). However, extensive effort in the mapping of phosphorylation sites in Cdc25A and the use of cells deficient in Chk2 or ATM indicate that many such sites are not required for Cdc25A-mediated G2-M checkpoint following DNA damage ( 20, 43– 45). Our data suggest that miR-21-mediated Cdc25A down-regulation facilitates the rapid establishment of the G2-M checkpoint following DNA damage. Interestingly, the elevated Cdc25A levels in unstressed miR-21 knockout cells do not appear to affect proliferation but profoundly affect cell cycle checkpoint and progression following stress (DNA damage or serum starvation), suggesting the importance of fully inactivating Cdc25A under these conditions. These conditional phenotypes associated with miR-21 might be particularly relevant, as growth factor deprivation and DNA damage have been shown to play important roles in tumorigenesis ( 46). In addition, Cdc25A was recently found to be a target of miR-16 that participates in UV-induced DNA damage response ( 42). Taken together, these observations suggest that critical cell cycle regulators such as Cdc25A are subject to modulation by miRNAs.
Our data provide a novel mechanism of how miR-21 could potentially contribute to tumorigenesis by compromising cell cycle progression and DNA damage–induced checkpoint function under those conditions, which can lead to chromosomal instability that promotes tumorigenesis ( 47). The cell cycle is composed of highly regulated machinery; the precise coordination of a timely entry into and exit from various stages during normal cell cycle is crucial for maintaining normal cell division that entails faithful DNA replication and segregation. In addition, most, if not all, of the cells in the human body are constantly encountering endogenous or exogenous insults that can damage DNA, and proper activation of checkpoints and recovery from them is probably just as important in ensuring genome integrity. Altered expression of miR-21 can conceivably cause genomic instability and lead to oncogenesis by relaxing or tightening this engine driving cell cycle through Cdc25A-dependent activation of cyclin/cyclin-dependent kinase complexes and may also affect therapeutic responses. Similar to Chk2-deficient cells ( 48), miR-21 knockout cells exhibit compromised checkpoint and radioresistance. Given the complexity of the regulation of miRNA targets, much work remains to define and characterize miR-21 targets to better understand its biology in different tissues and cancer. Therefore, future work will determine whether miR-21 affects chromosomal stability following DNA damage and other aspects of tumor biology through novel targets.
Overexpression of Cdc25A and Cdc25B is correlated with more aggressive disease and poor prognosis in some cancer patients ( 19). The reasons for Cdc25A overexpression are still not clear. Our data offer reduced miR-21 expression as a plausible explanation of Cdc25A overexpression in perhaps a subset of colon cancers. Other factors such as overexpression of c-Myc and E2F or inactivation of glycogen synthase kinase-3β are likely to be involved ( 19, 49). It is established that Cdc25A activities are tightly regulated by multiple mechanisms during cell cycle, including inhibitory and activating phosphorylation, changes in intracellular localization, and interactions with other proteins ( 17). Given a central role of Cdc25A in regulating cell cycle progression, it is perhaps not surprising that additional mechanisms such as miR-16 ( 42) can fine-tune its activity or levels.
Lastly, miR-21 appears to regulate a distinct set of genes and have a limited role in regulating anticancer drug–induced apoptosis in colon cancer cells. The discrepancies in targets identified by different groups are perhaps not surprising, as miRNAs are known to regulate targets in a tissue-specific and cell type–specific manner ( 6). It is also possible that some of these targets are primarily regulated by miR-21 at the level of translation. The miR-21 targeted cells and the targeting vector established in this study should be very useful for further dissecting miR-21 biology.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Grant support: Flight Attendant Medical Research Institute; Alliance for Cancer Gene Therapy (J. Yu); NIH grants 1R01CA129829, U19-A1068021 (pilot project; J. Yu), CA106348, CA121105 (L. Zhang), CA127590, and U54 CA116867 (Z. Wang); and American Cancer Society grant RSG-07-156-01-CNE (L. Zhang). L. Zhang and Z. Wang are V scholars.
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.
We thank the other members of our laboratories and Dr. Edward V. Prochownik (University of Pittsburgh) for helpful discussion and comments, Hongtao Liu for technical assistance, and Dr. Jianhua Luo and the Microarray Core Facility at University of Pittsburgh School of Medicine for gene expression analysis.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
P. Wang and F. Zou contributed equally to this work.
Current address for F. Zou: Department of Biology, Sichuan University, Chengdu, People's Republic of China 610064.
- Received June 1, 2009.
- Revision received August 3, 2009.
- Accepted August 27, 2009.
- ©2009 American Association for Cancer Research.