MicroRNA-195 suppresses tumorigenicity and regulates G1/S transition of human hepatocellular carcinoma cells


  • Teng Xu,

    1. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ying Zhu,

    1. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yujuan Xiong,

    1. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences
    Search for more papers by this author
  • Yi-Yuan Ge,

    1. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences
    Search for more papers by this author
  • Jing-Ping Yun,

    1. State Key Laboratory of Oncology in Southern China
    2. Department of Pathology, Cancer Center, Sun Yat-sen University, Guangzhou, P.R. China
    Search for more papers by this author
  • Shi-Mei Zhuang

    Corresponding author
    1. Key Laboratory of Gene Engineering of the Ministry of Education, State Key Laboratory of Biocontrol, School of Life Sciences
    2. State Key Laboratory of Oncology in Southern China
    • Key Laboratory of Gene Engineering of the Ministry of Education, School of Life Sciences, Sun Yat-sen University, Xin Gang Xi Road 135#, Guangzhou 510275, P.R. China
    Search for more papers by this author
    • fax: +86-20-84112169

  • Potential conflict of interest: Nothing to report.


Growing evidence indicates that deregulation of microRNAs (miRNAs) contributes to tumorigenesis. Down-regulation of miR-195 has been observed in various types of cancers. However, the biological function of miR-195 is still largely unknown. In this study we aimed to elucidate the pathophysiologic role of miR-195. Our results showed that miR-195 expression was significantly reduced in as high as 85.7% of hepatocellular carcinoma (HCC) tissues and in all of the five HCC cell lines examined. Moreover, introduction of miR-195 dramatically suppressed the ability of HCC and colorectal carcinoma cells to form colonies in vitro and to develop tumors in nude mice. Furthermore, ectopic expression of miR-195 blocked G1/S transition, whereas inhibition of miR-195 promoted cell cycle progression. Subsequent investigation characterized multiple G1/S transition-related molecules, including cyclin D1, CDK6, and E2F3, as direct targets of miR-195. Silencing of cyclin D1, CDK6, or E2F3 phenocopied the effect of miR-195, whereas overexpression of these proteins attenuated miR-195-induced G1 arrest. In addition, miR-195 significantly repressed the phosphorylation of Rb as well as the transactivation of downstream target genes of E2F. These results imply that miR-195 may block the G1/S transition by repressing Rb-E2F signaling through targeting multiple molecules, including cyclin D1, CDK6, and E2F3. Conclusion: Our data highlight an important role of miR-195 in cell cycle control and in the molecular etiology of HCC, and implicate the potential application of miR-195 in cancer therapy. (HEPATOLOGY 2009.)

It has been well demonstrated that aberration of protein-coding genes plays a crucial role in the development of human cancers. Recently, growing evidence has indicated that deregulation in noncoding genes, particularly microRNAs (miRNAs), also contributes to tumorigenesis.1, 2

miRNAs belong to a class of short, highly conserved noncoding RNAs known to suppress the expression of protein-coding genes through imperfect base-pairing with the 3′-untranslated region (3′ UTR) of target messenger (m)RNA.3 It is predicted that one-third of protein-coding genes in humans are regulated by miRNAs.4 Indeed, miRNAs have been implicated in the control of various biological processes, such as development,5 cell proliferation, apoptosis,6 and differentiation.7 Additionally, more than half of the annotated human miRNA genes are located in chromosomal regions displaying frequent amplification, deletion, or translocation in human cancers.8 Furthermore, deregulation of miRNAs has been observed in different types of human malignancies1, 9–14 and is also associated with the clinical outcome of cancer patients.11, 12, 15 Accumulating evidence suggests that miRNAs may act as oncogenes or tumor suppressor genes.2

Using miRNA microarray, we and others have previously shown down-regulation of the miR-15/16/195 family members in hepatocellular carcinoma (HCC) as well as other human neoplasms.9, 12–14 It was recently reported that ectopic expression of this family's members results in down-regulation of cell-cycle-related transcripts, including CDK6, CCNE1, Cdc25A, and Chk1, and causes an accumulation of G1 cells.16 However, the signaling pathways regulated by the miR-15/16/195 family and the exact role of this family in tumorigenesis are still largely unknown.

It has been demonstrated that members of an miRNA family usually play overlapping, if not identical, roles,17–21 because they share common binding sequences for target recognition. Therefore, miR-195, which, among the miR-15/16/195 family members, displayed the most significantly reduced expression in HCC in our previous study,9 was chosen as representative to clarify the pathophysiologic role and the relevance of the miR-15/16/195 family to tumor biology. We showed that down-regulation of miR-195 occurred very frequently in HCC tissues and cell lines. Furthermore, ectopic expression of miR-195 dramatically suppressed HCC cells to form colonies in vitro and to develop tumors in nude mice. Moreover, both gain- and loss-of-function studies revealed that miR-195 could block the G1/S transition. Multiple G1/S transition-related molecules, including cyclin D1, CDK6, and E2F3, were further characterized as direct functional targets of miR-195. These findings suggest an important role of miR-195 in cell cycle control and in the molecular etiology of HCC, with implications for cancer therapy.


FACS, fluorescence-activated cell sorting; HCC, hepatocellular carcinoma; miRNA, microRNA; NC, negative control; ppRb, phosphorylated pRb; RT-PCR, reverse-transcriptase polymerase chain reaction; 3′ UTR, 3′-untranslated region.

Materials and Methods

Cell Lines and Tissue Specimens.

Detailed information about the HCC cell lines and tissue specimens is described in the Supporting Materials and Methods. The relevant characteristics of the studied subjects are shown in Supporting Table 1.

RNA Oligoribonucleotides.

miRNA duplexes corresponding to miR-195, miR-16, and miR-497 were designed as described.22 The small interfering (si)RNAs targeting human cyclin D1 (GenBank Access. No. NM_053056), CDK6 (NM_001259), E2F3 (NM_001949), and c-Myc (NM_002467) transcripts were designated siCCND1, siCDK6, siE2F3, and siMyc, respectively. The negative control RNA duplex (denoted NC), designed based on cel-miR-67, has been demonstrated to have minimal sequence identity with human miRNAs (Dharmacon, Lafayette, CO). For the in vivo tumorigenicity assay, all pyrimidine nucleotides in the NC or miR-195 duplex were replaced by their 2′-O-methyl analogs to improve RNA stability. The anti-miR-195, with a sequence complementary to the mature miR-195, was a 2′-O-methyl-modified oligoribonucleotide. The anti-miR-C was used as a negative control for anti-miR-195. All RNA oligoribonucleotides (Supporting Table 2) were purchased from Genepharma (Shanghai, China).

Semiquantitative Reverse-Transcriptase Polymerase Chain Reaction (RT-PCR), Northern and Western Blot.

Primers used for semiquantitative RT-PCR and probes used for Northern blot analysis are listed in Supporting Table 2.

Northern blot was conducted as reported.23 We first optimized 55°C as the proper hybridization temperature to distinguish miR-195 from miR-16, which shares the highest sequence similarity with miR-195 (Supporting Fig. 1).

The antibodies used for Western blot are described in the Supporting Materials and Methods.

RNA and protein bands were quantified using GeneTools software (version 3.03; SynGene, Cambridge, UK).

Cell Transfection.

Reverse transfection of RNA duplex was performed using Lipofectamine-RNAiMAX (Invitrogen, Carlsbad, CA). All RNA transfections were performed at a final concentration of 50 nM unless otherwise indicated. The transfection efficiency, determined by a fluorescein amidite (FAM)-conjugated siRNA (nonhomologous to any human genome sequences) and fluorescence-activated cell sorting (FACS) analysis, was about 79% for MHCC-97L and 71% for HCT-116 cells (data not shown). Transfection of plasmid DNA was performed using calcium phosphate precipitation for 293T cells and Lipofectamine 2000 (Invitrogen) for other cell lines.

Analysis of Clonogenicity In Vitro and Tumorigenicity in Nude Mice.

Analyses for clonogenicity in low cell density, anchorage-independent cell growth, and tumorigenicity were performed as described in the Supporting Materials and Methods.

Analysis of Cell Proliferation.

Cell proliferation was determined by Alamar Blue assay (AbD Serotec, Oxford, UK) and bromodeoxyuridine (BrdU) incorporation assay as described.9, 23

Cell Cycle Analysis.

Except for rescue experiments, all cell cycle analyses were performed using a detergent-containing hypotonic solution (Krishan's reagent) and FACS (Becton Dickinson, San Jose, CA), as reported.24 Nuclear debris and overlapping nuclei were gated out unless stated otherwise.

For rescue assay, the DNA content of living cells was measured by FACS using Hoechst 33342, as described in the Supporting Materials and Methods.

Vector Construction and Luciferase Reporter Assay.

The expression vectors pc3-miR-195, pc3-gab-CCND1, pc3-gab-CDK6, and pc3-gab-E2F3 were generated as described in the Supporting Materials and Methods. To create a luciferase reporter construct, 3′ UTR fragment containing putative binding sites for miR-195 was inserted downstream of the stop codon of firefly luciferase in pGL3cm as described.9 Mutant 3′ UTR, which carried a mutated sequence in the complementary site for the seed region of miR-195, was generated using the fusion PCR method.

293T cells grown in a 48-well plate were cotransfected with 200 ng of either pcDNA3.0 or pc3-miR-195, 10 ng of firefly luciferase reporter comprising wildtype or mutant 3′ UTR of target gene, and 2 ng of pRL-TK (Promega, Madison, WI). The luciferase assay was performed as reported.9


Paraffin-embedded, formalin-fixed tissues were immunostained for E2F3, cyclin D1, and CDK6 proteins, as described in the Supporting Materials and Methods.

Statistical Analysis.

Data are expressed as the mean ± SEM from at least three separate experiments. Unless otherwise noted, the differences between groups were analyzed using Student's t test when only two groups, or assessed by one-way analysis of variance (ANOVA) when more than two groups were compared. All tests performed were two-sided. Differences were considered statistically significant at P < 0.05.


Expression of miR-195 Is Frequently Reduced in HCC Tissues.

In a previous microarray analysis, we found that miR-195 was down-regulated in HCC tissues compared with normal liver tissues.9 To further confirm and extend this finding, we analyzed the expression of miR-195 in 14 paired HCC and adjacent noncancerous liver tissues by Northern blot. Compared with their noncancerous counterparts, significant down-regulation of miR-195 was observed in 85.7% (12/14) of HCC samples (Fig. 1). Furthermore, 9 out of 14 (64.3%) HCC displayed more than 50% reduction in the miR-195 level. Down-regulation of miR-195 was also found in all five HCC cell lines examined (Supporting Fig. 2). These results suggest that reduced miR-195 expression is a frequent event in human HCC and may be involved in hepatocarcinogenesis.

Figure 1.

Analysis of miR-195 expression in paired HCC and adjacent nontumor tissues by Northern blot. T, HCC tissue; N, adjacent noncancerous tissue. The same membrane was hybridized sequentially with miR-195 and a U6 probe. The intensity for each band, representing gene expression level, was densitometrically quantified. miR-195 level was normalized by the intensity of U6 in each sample. The value under each pair of samples (T/N) indicates the fold change of miR-195 level in HCC tissue, relative to that in adjacent nontumor tissue.

miR-195 Suppresses Clonogenicity In Vitro and Tumor Growth In Vivo.

To examine the potential role of miR-195 in tumorigenesis, we first evaluated the effect of miR-195 on the growth and clonogenicity of cancer cells. Two HCC cell lines (MHCC-97L, Hep3B) were transfected with NC or miR-195 duplex, and then allowed to grow at very low density. Notably, miR-195–transfected cells displayed obviously fewer and smaller colonies compared with NC-transfectants (Fig. 2A), suggesting a growth-inhibitory role of miR-195. To test whether this observation could be generalized to other solid tumor cells, miR-195 was transfected into colorectal carcinoma (HCT-116) and osteosarcoma (U-2OS) cells. Interestingly, both of them also exhibited dramatically decreased colonies after miR-195 expression (Fig. 2A), indicating a ubiquitous role of miR-195 in cell growth. This growth-inhibitory effect of miR-195 was also confirmed by Alamar Blue assay (Supporting Fig. 3).

Figure 2.

miR-195 suppresses clonogenicity in vitro and tumor growth in vivo. (A) Colony formation at very low cell density. (B,C) Tumor formation in nude mice. Representative photographs (B) and the curve of tumor growth (C) are shown. NC and miR-195 in Fig. 2B indicates the flanks injected with NC-transfected and miR-195-transfected cells, respectively. *P < 0.05, compared with NC-transfectants. Wilcoxon signed ranks test was used for the comparison of tumor volumes.

We further analyzed the effect of miR-195 on anchorage-independent cell growth and found that miR-195 dramatically repressed the capability of cancer cells to grow in soft agar, as shown by a significant reduction both in the size and the number of colonies (Supporting Fig. 4). The above findings were further confirmed by an in vivo xenograft model. NC- and miR-195-tranfected MHCC-97L cells were injected subcutaneously into either posterior flank of the same nude mice. The mice were followed for observation of xenograft growth for 4 weeks. We found that introduction of miR-195 into MHCC-97L cells led to a significant reduction in both the frequency of tumor formation (NC versus miR-195-transfectants, 5/5 versus 3/5) and the size of tumor volume (Fig. 2B,C, upper panel). This suppression on tumorigenicity was also reproducible in HCT-116 xenografts. Tumors appeared in the sites injected with NC-transfected HCT-116 in 56% (5/9) of mice, and became palpable 12-22 days after inoculation and grew to 146∼1,386 mm3 at the end of observation (4 weeks). In sharp contrast, no tumors were observed in the flanks injected with miR-195-transfectants through the observation period (Fig. 2B,C, lower panel).

Collectively, both in vitro and in vivo studies suggest that miR-195 significantly inhibits tumorigenicity.

Ectopic Expression of miR-195 Blocks G1/S Transition.

To explore the mechanisms underlying miR-195-suppressed tumor growth, predicted target genes of miR-195 were retrieved from the TargetScan database and their functions were categorized based on the Protein Lounge Pathway Database (http://www.proteinlounge.com). The enrichment of predicted targets in defined biological pathways was then evaluated by a chi-square test. Interestingly, we found a significant enrichment of cell cycle-related genes among the predicted targets (4.2-fold overrepresentation, P < 10−4; Supporting Table 3). To confirm this finding, target prediction for miR-195 was further performed using another algorithm, miRanda,25 with minimal score and free energy set as 90 and −17 cal/mol, respectively. A similar trend of enrichment in cell cycle-related transcripts was also observed (5.1-fold overrepresentation, P < 10−7; Supporting Table 3).

Such an enrichment prompted us to investigate the impact of miR-195 on cell cycle progression. The results revealed that enforced expression of miR-195 caused marked accumulation of a G1 population in different cell lines, whereas little effect on apoptosis was observed as evaluated by sub-G1 peak (Supporting Fig. 5). To further confirm this finding, we synchronized the transfected cells with nocodazole, which depolymerizes the microtubules and blocks cell cycle at G2/M-phase. Thirty-two hours after transfection, cells were incubated with nocodazole for an additional 16 hours before FACS analysis. Consistently, miR-195-transfected cells showed a substantial rise in the G1 population (Fig. 3A and Supporting Fig. 6A). Similar results were obtained when cells were transfected with miR-16 or miR-497 (Supporting Fig. 7), two other members of the miR-15/16/195 family, in agreement with the general concept of functional redundancy among paralogous family members.

Figure 3.

Cell cycle analysis by FACS. (A) Effect of miR-195 overexpression and knockdown of either cyclin D1, CDK6, or E2F3 on the G1-population. **P < 0.01; ***P < 0.001, compared with NC-transfectants. (B) Effect of miR-195 overexpression on S-phase entry upon serum stimulation. NC- or miR-195-transfectants were serum-deprived, followed by serum readdition 48 hours later and then harvest at indicated timepoints. The timepoint when serum was readded was set as 0 hour.

Next, a serum starvation-stimulation assay was employed to precisely clarify whether miR-195-induced accumulation of G1-cells resulted from a blocked G1/S transition. MHCC-97L, HCT-116, and U-2OS cells were transfected with NC or miR-195 for 24 hours, followed by synchronization with serum deprivation for 48 hours and then stimulation to enter S-phase by serum readdition. Interestingly, when a large proportion of NC-transfected cells already entered S-phase, most of the miR-195-transfectants still stayed in G1-phase after serum stimulation (Fig. 3B and Supporting Fig. 6B), suggesting that miR-195 blocks the G1/S transition. This contention was further supported by BrdU incorporation assay (Supporting Fig. 6C).

miR-195 Directly Inhibits the Expression of Cyclin D1, CDK6, and E2F3 through Their 3′ UTRs.

To unravel the mechanism by which miR-195 restrains the G1/S transition, we searched for positive regulators of a G1/S transition among the predicted targets of miR-195. Cyclin D1, CDK6, and E2F3 stood out as attractive candidates, because they are crucial components that initiate the inactivation of the Rb pathway and in turn the G1/S transition. We were able to map a single putative miR-195-binding site in the 3′ UTR of E2F3 and two sites in each 3′ UTR of cyclin D1 and CDK6 (Supporting Fig. 8). To validate whether these genes were direct targets of miR-195, a dual-luciferase reporter system was first employed. Firefly luciferase reporter containing wildtype or mutant 3′ UTR of the target gene was cotransfected with Renilla luciferase reporter and either pcDNA3.0 or pc3-miR-195. Coexpression of pc3-miR-195 significantly suppressed firefly luciferase activity of the reporter with wildtype 3′ UTR but not that of the mutant reporter (Fig. 4A), indicating that miR-195 may suppress gene expression through its binding sequences at the 3′ UTR of target genes.

Figure 4.

Cyclin D1, CDK6, and E2F3 are direct targets of miR-195. (A) Analysis of luciferase activity. Cells were cotransfected with pc3-miR-195 or pcDNA3.0, firefly luciferase reporter containing either wildtype or mutant 3′ UTR (indicated as Wt or Mut on the X axis), and Renilla luciferase-expressing construct (as an internal control to correct for the differences in both transfection and harvest efficiencies). The firefly luciferase activity of each sample was normalized to the Renilla luciferase activity. The normalized luciferase activity of pcDNA3.0-transfectant in each experiment was set as relative luciferase activity 1; therefore, no error bar is shown for pcDNA3.0-transfectant. *P < 0.05; ***P < 0.001, compared with pcDNA3.0-transfectants. (B) Effect of miR-195 or siRNA on the expression of endogenous cyclin D1, CDK6, and E2F3. Forty-eight hours after transfection with the indicated RNA duplex, endogenous mRNA and protein levels were assayed by RT-PCR and Western blot, respectively. β-Actin and hPRT, internal control. (C) Immunohistochemistry staining of E2F3 in paired HCC and adjacent noncancerous tissues. Samples with the lowest (sample 652, left panel) and the highest (sample 729, right panel) level of miR-195 are presented. The fold changes of miR-195 expression in HCC 652 and 729 (as measured by the T/N ratio in Fig. 1) were 0.06 and 1.22, respectively. Images were captured at 400×.

The effect of miR-195 on endogenous expressions of cyclin D1, CDK6, and E2F3 was subsequently examined. Transfection of miR-195 induced an obvious decrease in all three proteins in both MHCC-97L and U-2OS cells, and in CDK6 and E2F3 in HCT-116. In addition, miR-195 overexpression caused a reduced mRNA level in CDK6 but not in cyclin D1 and E2F3 (Fig. 4B and Supporting Fig. 6D).

We next examined the correlation between the decreased expression of miR-195 and the overexpression of E2F3, cyclin D1, and CDK6, the targets of miR-195, in HCC tissues. E2F3, cyclin D1, and CDK6 were analyzed by immunohistochemistry in the same set of specimens shown in Fig. 1. Notably, E2F3 was highly expressed in 71.4% (10/14) of HCC tissues, but undetectable in all of the corresponding noncancerous samples (Supporting Table 1 and Fig. 4C). Importantly, 10 out of 12 HCC displaying a striking decrease in the miR-195 level revealed enhanced expression of E2F3, whereas those two HCC without an obvious reduction of miR-195 expression showed undetectable E2F3 staining (cases 729 and 727; Supporting Table 1; Figs. 1, 4C). Consistently, statistical analysis revealed that the relative miR-195 level (as measured by the tumor/nontumor [T/N] ratio) was significantly lower in HCC with E2F3 overexpression (T/N = 0.37 ± 0.06) than those without E2F3 staining (T/N = 0.74 ± 0.20; P < 0.05 by t test). These data suggest that HCC tissues with lower miR-195 levels predispose to E2F3 overexpression.

We also found that cyclin D1 and CDK6 were overexpressed in 21.4% (3/14) and 85.7% (12/14) of samples, respectively (Supporting Table 1). Furthermore, a trend of correlation between the decreased miR-195 level and the cyclin D1 or CDK6 overexpression was also observed. The relative miR-195 level in HCC with cyclin D1 or CDK6 overexpression was about 2-fold lower than those without enhanced expression of cyclin D1 or CDK6 (T/N = 0.24 ± 0.01 versus 0.53 ± 0.09 for cyclin D1; T/N = 0.41 ± 0.06 versus 0.85 ± 0.38 for CDK6), although the correlation did not reach statistical significance (P = 0.14 for cyclin D1; P = 0.05 for CDK6 by t test), which might be due to the limited sample size after stratification based on the expression status of cyclin D1 or CDK6 (n ≤ 3 in the minor group).

Taken together, these data imply that miR-195 may attenuate the expression of cyclin D1, CDK6, and E2F3 by directly targeting their 3′ UTRs.

Cyclin D1, CDK6, and E2F3 Are Involved in miR-195–Regulated G1/S Transition.

To explore the role of cyclin D1, CDK6, and E2F3 in miR-195-regulated G1/S transition, we examined whether knockdown of these genes may mimic the effect of miR-195 overexpression. Cells were transfected with siRNA duplex targeting either cyclin D1, CDK6, or E2F3, which resulted in a significant reduction in both mRNA and protein levels of the respective genes (Fig. 4B and Supporting Fig. 6D). Markedly, except for knockdown of cyclin D1 in MHCC-97L, silencing of either target genes led to substantial arrest of cell cycle in G1-phase (Fig. 3A and Supporting Fig. 6A), phenocopying the outcome of miR-195 expression. Interestingly, compared with miR-195 overexpression, inhibition of cyclin D1, CDK6, or E2F3 alone caused a less extent of G1-arrest, implying that miR-195 may repress the G1/S transition through synergistically targeting multiple targets. We subsequently examined whether cyclin D1, CDK6, and E2F3 could counteract the effect of miR-195. Notably, overexpression of these proteins partially rescued miR-195-induced G1-arrest in MHCC-97L and U-2OS cells (Fig. 5A and data not shown).

Figure 5.

Rb-E2F signaling is involved in miR-195-regulated G1/S transition. (A) Expression of cyclin D1, CDK6, and E2F3 partially rescues exogenous miR-195-induced G1-arrest. Cells were first transfected with NC or miR-195 for 24 hours, followed by transfection with empty pc3-gab vector (indicated as EGFP) or the mixture of pc3-gab-CCND1, pc3-gab-CDK6, and pc3-gab-E2F3 vectors (indicated as Mixture), which encoded the entire coding sequence of each gene but lacked the 3′ UTR. **P < 0.01; ***P < 0.001. (B) Phosphorylation of Rb protein is inhibited by miR-195. Forty-eight hours after transfection with the indicated RNA duplex, cells were subjected to Western blot analysis. pRb, total Rb protein; ppRb, phosphorylated pRb. β-Actin, internal control. (C) Transactivation of E2F3 target genes is impeded by miR-195. NC-transfectants or miR-195-transfectants were serum-deprived, followed by serum readdition 48 hours later and then harvested at the indicated timepoints. The timepoint when serum was readded was set as 0 hour. Gene expression was examined by RT-PCR. hPRT, internal control.

Cyclin D1 and CDK6 are crucial molecules that initiate the phosphorylation of Rb, which results in the release of E2F and in turn the transactivation of genes required for S-phase entry. We therefore explored whether miR-195 could attenuate these events. Interestingly, ectopic delivery of miR-195 caused a prominent decrease of phosphorylated-Rb protein (ppRb) (Fig. 5B and Supporting Fig. 6E). Furthermore, the mRNA levels of downstream targets of E2F, including Mcm3, DHFR, TK1, Cdc6, cyclin E1, and cyclin A2 genes, were significantly lower in miR-195-transfected cells than in NC-transfectants (Fig. 5C and Supporting Fig. 6F).

These data suggest that miR-195 may inhibit the G1/S transition by suppressing multiple targets including cyclin D1, CDK6, and E2F3, and in turn attenuating the transcription of S-phase genes.

Inhibition of miR-195 Promotes G1/S Transition.

To confirm the results from the gain-of-function study, we performed loss-of-function analysis using anti-miR-195, an inhibitor of miR-195. Transfection of anti-miR-195 dramatically counteracted the G1-arrest triggered by exogenous miR-195 (Fig. 6A). The outcome of antagonism on endogenous miR-195 was further investigated. Compared with anti-miR-C transfection, anti-miR-195 induced a statistically significant decrease of G1-cells (50.4% versus 60.3%, P = 0.003, Fig. 6B), although the extent of antagonism was not as evident as that on exogenous miR-195, which might be explained by the low basal level of miR-195 in cancer cells (Supporting Fig. 2).

Figure 6.

Inhibition of miR-195 promotes G1/S transition. (A) miR-195 inhibitor counteracts the G1-arrest triggered by exogenous miR-195. HCT-116 cells were cotransfected with miR-195 duplex (20 nM) together with anti-miR-195 or anti-miR-C (100 nM). Nocodazole (40 ng/mL) was added 32 hours after transfection and cells were cultured for an additional 16 hours, followed by FACS analysis. ***P < 0.001, compared with NC-transfectants. (B) Effect of miR-195 inhibitor in HCT116. Cells were transfected with anti-miR-NC (black line) or anti-miR-195 (gray histogram) for 72 hours, followed by addition of nocodazole and culture for an additional 4 hours before FACS analysis. P was obtained by comparing G1 proportions between anti-miR-NC-transfectants and anti-miR-195-transfectants. (C,D) Effect of miR-195 inhibitor in c-Myc-knockdown HCT116. Cells were cotransfected with the indicated siRNA duplex (5 nM) and miRNA inhibitor (200 nM), followed by FACS (C) and Western blot (D) analysis. For FACS analysis, nocodazole (40 ng/mL) was added 32 hours after transfection and cells were cultured for an additional 16 hours before harvest. For Western blot, cells were collected 48 hours after transfection.

It has been shown that c-Myc-silencing can induce G1-arrest by reducing the expression of cell cycle-related molecules,26–28 including cyclin D1 and E2F3, the targets of miR-195 identified in this study. We therefore attempted to further confirm the effect of anti-miR-195 on endogenous miR-195 using a c-Myc-knockdown model, in which the increased G1 population and the reduced expression of cyclin D1 and E2F3 may make the antagonism of anti-miR-195 more pronounced. HCT-116, a cell line with a high level of endogenous c-Myc, was transfected with siMyc, which significantly repressed c-Myc expression (Supporting Fig. 9). As expected, compared with NC-transfected cells, siMyc-transfectants displayed an increased proportion of G1-cells (Fig. 6C) and a reduced level of cyclin D1, CDK6, E2F3, and ppRb (Fig. 6D). Surprisingly, transfection of anti-miR-195 not only significantly enhanced the levels of cyclin D1, CDK6, E2F3, and ppRb, but also dramatically relieved the G1-arrest in c-Myc-knockdown cells (Fig. 6C,D).

These findings, together with the data from gain-of-function experiments, point to a role of miR-195 in cell cycle regulation.


It is well demonstrated that a defect in cell cycle control is an essential step during carcinogenesis. Therefore, it is reasonable to expect that deregulation of cell cycle-related miRNAs may facilitate tumorigenesis. In this study we showed that miR-195 was frequently down-regulated in both HCC tissues and cell lines. Ectopic expression of miR-195 suppressed HCC and colorectal carcinoma cells to form colonies in vitro and to develop tumors in vivo. Moreover, gain- and loss-of-function studies revealed that miR-195 could block the G1/S transition by repressing the Rb-E2F signaling through multiple targets, including cyclin D1, CDK6, and E2F3. We propose that reduced expression of miR-195 may disrupt cell cycle control, in turn promote cell proliferation, and consequently facilitate the development of cancers like HCC.

The underlying mechanism responsible for decreased expression of miR-195 in HCC is still unknown. Notably, miR-195 is located at chromosome 17p13.1, a region that is frequently deleted in human cancers, especially HCC. In addition, we were able to computationally map a CpG-enriched region upstream of the miR-195 gene. Therefore, allelic loss and promoter hypermethylation might partially account for the reduced miR-195 expression in cancers.

The miR-15/16/195 family has been shown to promote apoptosis in leukemia cells,29 whereas this effect has not been observed in the cell models used in the present study and previous reports.16, 30 This discrepancy is likely due to the difference in cell context, and suggests that altered expression of this miRNA family may have diverse effects in tumor cells. Nevertheless, the similar effects of miR-195 in cell lines derived from different solid cancers implicate its fundamental role in cell cycle regulation and tumorigenesis.

Intriguingly, the promoting effect of anti-miR-195 on cell cycle progression was exaggerated after c-Myc-knockdown (Fig. 6B,C). We speculate that c-Myc-silencing may provide a more proper condition for examining the effect of miR-195 inhibition. For example, the anti-miR-195-induced increase in the expression of miR-195 targets may become more evident when these targets are expressed at low levels after c-Myc-knockdown. Besides, a recent study showed that c-Myc could repress the transcription of miR-195/497 and miR-15a/16 polycistrons in lymphoma cells.31 However, we could not find an obvious elevation in either miR-195 and miR-16 levels after c-Myc-silencing (Supporting Fig. 10).

Although miR-195 effectively suppressed cyclin D1 expression in diverse cell lines examined in our and other studies,30 we noticed that none of miR-195, miR-16, and miR-497 were able to reduce the level of cyclin D1 in HCT-116 cells (Fig. 4B and Supporting Fig. 11). It is becoming increasingly clear that miRNA-mediated suppression of gene expression is subjected to intense regulation. For example, the repression of CAT-1 by miR-122 in HCC cells is relieved under stress conditions by binding of HuR to the 3′ UTR of CAT-1.32 Furthermore, it has been shown that Dnd1 protein can inhibit the access of miRNA to target mRNA, and therefore block the regulation function of miRNA on gene expression.33 It will be interesting to characterize how the interaction between miR-195 and the mRNA of cyclin D1 is controlled in HCT-116.

While this article was in preparation, Liu et al.30 reported that the miR-15/16/195 family directly targeted cyclin D1 and CDK6, and forced expression of either miR-16 or miR-195 increased G1-cell proportion in a lung cancer cell line. In the present study, we confirmed and extended their findings in other cell models. We provided evidence to support the potential tumor suppressive activity of miR-195, further defined miR-195 as an inhibitor of G1/S transition, and characterized E2F3, a transcription factor downstream of Rb, as another target of miR-195. It is well known that the Rb pathway acts as a master checkpoint in cell cycle progression. The integration of miR-195 in the Rb-E2F pathway further demonstrates that cell cycle is regulated by elaborate mechanisms, to prevent uncontrolled cell proliferation. Our finding that miR-195 targets molecules both upstream (e.g., cyclin D1, CDK6) and downstream (e.g., E2F3) of Rb, an archetypal tumor suppressor, provides new insight into a fail-safe mechanism, which should guard the cells from runaway proliferation in the case of aberrant Rb-inactivation, which is frequently found in human HCC.34

In summary, we investigated the potential role of miR-195 in cell cycle control and in tumorigenesis. Our data suggest an important role of miR-195 in the molecular etiology of cancer and implicate its potential application in cancer therapy.