RNA interference is the process by which double-stranded RNA directs sequence-specific degradation of mRNA. It has recently been shown that RNA interference can be triggered by 21-nucleotide duplexes of small interfering RNAs (siRNAs) in both cultured mammalian cells and adult mice. We hypothesize that siRNA can be used to specifically target oncogene overexpression in a therapeutic manner. Here, we show that overexpression of the oncogene cyclin E can be suppressed by up to 90% in hepatocellular carcinoma (HCC) cell lines by siRNA targeted on the coding region of cyclin E. We also find that depletion of cyclin E in this manner promotes apoptosis of HCC cells and blocks cell proliferation. Finally, we show that the siRNA oligos inhibits HCC tumor growth in nude mice. Thus, this study demonstrates the therapeutic potential of siRNA on the treatment of HCC by targeting overexpressed oncogenes such as cyclin E. Our results also indicate that cyclin E, which is overexpressed in 70% of HCCs, may serve as a novel therapeutic target.
RNAi3 is the process whereby dsRNA results in the rapid destruction of mRNA containing the identical sequence as the dsRNA. The mediators of RNAi are 21- and 22-nt siRNAs generated by RNase III cleavage from longer dsRNA (1, 2). Delivery of dsRNA has been shown to knock down the expression of specific proteins in insect cell lines. However, in most mammalian cells, it has not been possible to exert potent and specific effects by applying dsRNA (>38 bp) because long dsRNA also induces a nonspecific inhibitory response resulting from the IFN pathway (3). Recently, it has been shown that delivery of 21-nt siRNA specifically suppressed expression of endogenous and heterologous genes in different mammalian culture cells and mice (4, 5, 6, 7). We sought to use siRNA targeting a common oncogene to determine whether this technique can be used to specifically inhibit oncogene overexpression and whether this inhibition results in antitumor effects.
HCC is the third leading cause of cancer death worldwide, with an estimated 564,000 new cases and almost as many deaths in 2000 (8, 9). Currently, there is no effective therapy for the vast majority of HCC patients. Therefore, the understanding of the molecular mechanisms involved in HCC formation and progression become critical to developing more effective treatments for HCC. Genetic analyses have revealed that one of the commonly altered genes in HCC is the cell cycle regulator cyclin E (10). Cyclin E is believed to control G1-S-phase progression. By associating with cyclin-dependent kinase Cdk2 and activating its kinase activity shortly before entry of cells into the S phase (11, 12), its expression near the G1-S-phase transition is thought to be critical for the initiation of DNA replication and duplication of the centrosomes. The timely appearance and disappearance of cyclin E is crucial: excessive activity of the cyclin E-Cdk2 complex drives cells to copy their DNA prematurely, resulting in genome instability (13) and carcinogenesis (14). In fact, clinical studies have indicated that cyclin E plays an important role in HCC formation and progression. Overexpression of cyclin E was found in ∼70% of HCC patients, which correlated with the poor prognosis of those patients (15). We choose to study the applicability of RNAi as therapy against cyclin E overexpression in HCC by in vitro and in vivo experiments.
MATERIALS AND METHODS
HCC cell lines, Hep3B, HepG2 and SNU449 were obtained from American Type Culture Collection (Manassas, VA) and HuH7 was a generous gift of Dr. Patricia Marion (Stanford University, Stanford, CA). Hep3B and HepG2 are cyclin E-overexpressed lines, and HuH7 is a cyclin E-nonoverexpressed line (16). The cells were grown in DMEM (Invitrogen) supplemented with 10% fetal bovine serum. All cells were maintained in a humidified 37°C incubator with 5% CO2.
Transfection with siRNA Oligos.
The siRNA oligos were synthesized by Dharmacon Research, Inc. The siRNA oligos corresponded to nts 361–382 of the human cyclin E coding region (GenBank accession no. XM_049430). The indicated HCC cells (1 × 105/well) were transfected with siRNA oligos (0.3 μg/well) in 6-well plates using Oligofectamine reagent (Invitrogen) following the manufacturer’s protocol.
Cell Cycle Analysis.
Standard fluorescence-activated cell sorter analysis was used to determine apoptosis of the cells or the distribution of cells in cell cycle. Briefly, the cells were transfected with cyclin E siRNA or other agents. Adherent cells were then collected by trypsinization and combined with cells floating in the medium. The apoptotic cells were assessed by flow cytometric detection of sub-G1 DNA content after being stained with propidium iodide.
Apoptotic cells were confirmed using the In Situ Cell Death Detection kit from Roche (Mannheim, Germany), following the manufacture’s instruction. The apoptotic cells (red staining) were counted under a microscope. The apoptosis index was defined by the percentage of red cells among the total cells of each sample. Six fields with >100 cells in each were randomly counted for each sample.
Thirty h after transfection of siRNA, the cells were split into 4-well chamber slides and incubated with culture medium containing BrdU for 4 h or 20 h. BrdU staining were performed using Zymed BrdU labeling kit (Zymed, San Francisco, CA) following the manufacturer’s protocol.
Forty-four h after transfection, cells were lysed, as indicated, into mammalian cell lysis buffer [20 mm Tris-HCl (pH 7.5), 150 mm NaCl, 0.5% NP40, 1 mm EDTA, 1 mm EGTA, 1 mm DTT) with 5 mm sodium fluoride, 1 mm sodium orthovanadate, 1 mm phenylmethyl sulfonylflouride, 2 mg/ml aprotinin, and 2 mg/ml leupeptin]. After centrifugation at 4°C (14,000 rpm, 15 min), lysates (20 μg) were analyzed by immunoblotting. Anticyclin E polyclonal antibody (C-19) was from Santa Cruz Biotechnology, and antiactin antibody (Ab-1) was from Oncogene Research (Boston, MA). Images were quantitated by NIH-Image software.
Colony Formation Assay in Soft Agarose.
The standard colony formation assay was used (17). Briefly, HCC cell line Hep3B was transfected with siRNA oligos targeting cyclin E or LacZ. Two days after the transfection, the cells (1 × 103 cells/well) were plated in 24-well plates in culture medium containing 0.35% agarose overlying a 0.7% agarose bottom layer and cultured at 37°C with 5% CO2. Five weeks later, the top layer of the culture was stained with p-iodonitrotetrazolium (1 mg/ml). Colonies >100 μm in diameter were counted.
Ex Vivo Tumor Inhibition.
Hep3B cells were transfected with or without siRNA oligos. Forty-four h after transfection, 2 × 106 cells were s.c. injected into nude mice (nu/nu, 8–9 weeks of age; Harlan Sprague Dawley, Madison, WI). The volume of the resulting tumor was measured weekly. The difference in tumor volume between treated groups was compared for statistic significance using the unpaired, two-tailed t test.
Suppression of Cyclin E Overexpression in HCC by RNAi.
Cyclin E is commonly overexpressed in HCC and correlated with poor prognosis of those patients (10). To address if cyclin E could serve as a therapeutic target for this cancer, we used siRNA oligos to deplete cyclin E expression in HCC cells. We designed one pair of siRNA targeting the coding region of cyclin E (Fig. 1,A). The siRNA oligos were transfected into three cyclin E overexpressing HCC cell lines, Hep3B, HepG2, and SNU449. Forty-four h after transfection, the protein lysates were harvested and analyzed by anticyclin E Western blotting. We found that cyclin E expression levels were suppressed by up to 90% in all three cell lines (Fig. 1,B, Lanes 2, 8, and 11). This suppression was also detected as early as 24 h after transfection. These data indicated that the siRNA could effectively suppress cyclin E overexpression. The inhibitory effect of the cyclin E siRNA was shown to be specific because a control oligo-targeting LacZ gene had no effect on cyclin E expression levels. In addition, siRNA oligos did not cause a nonspecific down-regulation of gene expression, as demonstrated by the actin control (Fig. 1 B). The effects of the siRNA were durable as we observed the suppression even 5 days after transfection (data not shown).
Induction of Apoptosis and Inhibition of DNA Synthesis in HCC Cells by RNAi Depletion of Cyclin E.
Three days after the transfection of cyclin E siRNA, we observed that cells with cyclin E overexpression shrank, rounded up, and detached from plates, suggesting apoptosis had occurred. In contrast, the lacZ siRNA control group remained attached on the dishes and showed normal morphology. To determine whether depletion of cyclin E promotes tumor cell death, flow cytometry was performed after transfection of siRNA oligos into Hep3B, HepG2, and SNU449. The cells were analyzed at different time points (72 and 96 h) after transfection and significant sub-G1 (apoptotic) populations were observed at 96 h (Fig. 2,A). We found that 16% of Hep3B cells underwent apoptosis after transfection of siRNA oligos versus 1% in the control group. The apoptotic population was even higher when HepG2 and SNU449 cells were tested (44 and 31%, respectively; Fig. 2,A). As a control, we transfected LacZ siRNA into these three overexpressing cells, and no significant apoptosis was observed (data not shown). To test if apoptosis can be triggered in cells without cyclin E overexpression, we also transfected cyclin E siRNA into HuH7, a cyclin E-nonoverexpressing HCC cell line. In contrast to cyclin E overexpressed cells, we did not observe significant apoptosis in HuH7 at 72 or 96 h after transfection (Fig. 2,B), although its cyclin E protein was effectively suppressed by cyclin E siRNA (Fig. 1,B, Lane 5). We also confirmed the apoptosis of Hep3B by TUNEL assay (Fig. 2 B). These data together suggested that depletion of cyclin E specifically triggered apoptosis in cyclin E-overexpressing cells.
To determine whether cyclin E overexpression is required for G1-S-phase transition and replication in cancer cells, we examined the rate of DNA synthesis of siRNA-treated Hep3B cells by BrdU incorporation assay. Thirty h after transfection, the cells were incubated with BrdU for 20 h followed by the BrdU staining. Fewer (51%) BrdU-stained cells (brown color) were observed in the cyclin E siRNA-treated group compared with those in mock treatment (66%) or the LacZ siRNA control group (65%), indicating that depletion of cyclin E-suppressed Hep3B DNA synthesis (Fig. 2, C and D). We also performed BrdU incorporation assay in both HepG2 and SNU449 cells and found similar results (Fig. 2, C and D). To address if depletion of cyclin E affects the cell distribution in the cell cycle, we performed flow cytometry 32 h after transfection. Our results showed a decreased cell population in S phase after cyclin E siRNA treatment in all three cell lines compared with control, as shown in Fig. 2 E. Taken together, our results indicate that the cyclin E siRNA exhibited a specific inhibitory effect on cyclin E-overexpressing HCC through promotion of apoptosis as well as inhibition of DNA synthesis.
Cyclin E siRNA Inhibits Cancer Cell Growth and Suppresses Tumor Formation in Nude Mice.
Next, we sought to determine whether cyclin E siRNA could serve as a therapeutic agent against HCC tumor formation in nude mice. We first tested the effect of cyclin E siRNA on the growth of Hep3B, HepG2, and SNU449 in cell culture. As shown in Fig. 3,A, after transfection, cyclin E siRNA significantly inhibited cell growth of all three cell lines as compared with LacZ siRNA control or mock treatment. The inhibitory effect on Hep3B cell growth was confirmed by soft agar assay as shown in Fig. 3 B. In contrast, we did not observe dramatic growth inhibitory effect on HuH7 cells (data not shown).
The antitumor activity of siRNA in Hep3B cells was additionally assessed using an ex vivo assay. Hep3B cells were first transfected in Petri dishes with or without cyclin E siRNA. The transfected cells were then injected into nude mice, and the growth of tumors was measured weekly. As shown in Fig. 4, cyclin E siRNA significantly suppressed tumor growth in mice as compared with control, indicating that targeting cyclin E by siRNA can exert a strong antitumor effect in vivo on cyclin E-overxpressing HCC.
Oncogene overexpression has been implicated in the development and progression of a variety of human cancers and, therefore, provides a potential target for cancer gene therapy (10). For years, research has focused on effective tools to specifically down-regulate oncogene overexpression such as antisense oligonucleotide strategy. However, there has been only limited success because of the lack of specificity and potency for this method (18). For example, screening of >20 oligomers is usually required before identifying one antisense that functions effectively, and the dose required for inhibiting gene expression is often not much different from doses that lead to nonselective toxicity.
The recent progress of RNAi techniques has demonstrated the potential to overcome those limitations. The selection of the targeting sequences of RNAi is less restricted, so the success rates of producing effective duplexes are higher (18). In addition, siRNA is dsRNA, which is more resistant to nuclease degradation as compared with antisense oligos and, therefore, have longer therapeutic effects than the antisense approaches. A recent study directly compared these two techniques and found that siRNA appeared to be quantitatively more efficient with more durable in cell culture (19). When tested in mice, only siRNA but not antisense oligos exhibited the inhibitory effects. This sequence specificity and its effective inhibitory effects have recently been successfully applied to suppress cancer cell growth induced by point mutation-activated Ras (20) or by BCR/ABL fusion gene (21). Our studies were designed to test whether siRNA can also been applied to selectively target cancer cells with overexpression of oncogenes, which is a more common cause of oncogene activation compared with point mutations and chromosome translocations.
Our results demonstrate that siRNA can effectively down-regulate oncogene overexpression with great specificity. As shown, the siRNA oligos could successfully deplete up to 90% of cyclin E in Hep3B, HepG2, and SNU449, three cell lines that each express at least 10-fold higher cyclin E than that in normal cells, indicating the potency of RNAi as a new strategy for cancer therapy. Also, the blockage of proliferation and induction of apoptosis in cultured cells and the tumor suppression effect in nude mice additionally support the effectiveness of this treatment. The induced apoptosis was only observed in cyclin E-overexpressing cells (Hep3B, HepG2, and SNU449) but not in cyclin E-nonoverexpressing cells (HuH7), and this specificity should increase the therapeutic index of RNAi-based therapies.
After cells were treated with cyclin E siRNA for 30 or 44 h, we observed a significant decrease of DNA replication in both cyclin E-overexpressed cells and the noncyclin E-overexpressed cell HuH7 (data not shown). The decreased replication rate as determined by BrdU incorporation is consistent with our flow cytometry results showing a decrease of the S-phase population 32 h after transfection. Cyclin E has been known to play an important role at G1-S transition, as well as DNA replication (11, 12). Therefore, the slowdown of DNA synthesis may have resulted from both the blockage of S-phase entry and DNA replication. Interestingly, in addition to the inhibition of DNA synthesis, depletion of cyclin E for 72 or 96 h triggered apoptosis in all three cyclin E-overexpressing cell lines that we tested. It is unclear how cyclin E depletion triggered apoptosis in cyclin E-overexpressing cells at this point. However, our flow cytometry data suggest that there might be different mechanisms to trigger apoptosis in different cell lines depending on the unique genetic context in the individual lines. For example, in HepG2 cells, depletion of cyclin E-induced apoptosis occurred after a dramatic increase of the G1 population and a decrease of S and G2 cells (preliminary data not shown), suggesting that the failure of S-phase entry may somehow trigger cell death in these cells. In contrast, Hep3B cells transfected with cyclin E siRNA underwent apoptosis after an increase of G2-M phase cells. It suggests that after cyclin E is depleted, the cells with incomplete DNA replication, instead of being blocked at intra-S phase, may enter G2-M phases. Thus, apoptosis could be triggered as a consequence of the cells trying to undergo mitosis in the presence of unreplicated DNA.
Interestingly, when apoptosis was induced (72 or 96 h) in Hep3B cells by depletion of cyclin E, we observed an increase of S-phase cells, which is not seen in the earlier (32 h) time point before apoptosis was triggered. We suspect that the increase of S-phase accompanying apoptosis may be an artifact from the incomplete DNA fragmentation of G2-M cells. Although all cells that complete apoptosis possess a unified DNA content as sub-G1, the cells in the middle of this process would likely contain a variable amount of DNA. For example, the G2-M cells with partially fragmented chromosomes would have DNA content between 2N and 4N, which is indistinguishable from the S-phase cells on flow cytometry profiles. In fact, for the same flow cytometry experiment, if we excluded the cells floating in the medium (apoptotic population), we then observed a decrease of S phase in the cyclin E-siRNA-treated group as compared with the control at 96 h (data not shown). Thus, this suggests that this increase of S phase at 96 h might result from the incomplete DNA fragmentation of G2-M apoptotic cells.
In addition to demonstrating the use of RNAi in cancer gene therapy, our results also indicate that cyclin E might serve as a novel therapeutic target for HCCs with cyclin E overexpression. Cyclin E has been suggested as an attractive target for molecular therapeutics both because it is overexpressed in a significant fraction of human tumors, including HCC (10), and because this overexpression is implicated in tumor formation (14). It has been postulated that interference with cyclin E expression or function could inhibit the neoplastic growth of a wide variety of cancers (10). However, to date, there has been no report to experimentally support this theory. The hurdles may have resulted from the lack of approaches to effectively deplete cyclin E in cancer cells. Here, we showed that when overexpressed cyclin E was depleted in HCC, it resulted in multiple antitumor effects in cancer cells such as blockage of DNA replication and induction of apoptosis in vitro and reduction of tumor growth in vivo. These results suggest that cyclin E overexpression may be essential for maintaining cell proliferation, as well as cell survival in cyclin E-overexpressing HCC. This is the first study, to our knowledge, showing that targeting cyclin E overexpression is a potential effective approach to treating cyclin E-overexpressing HCC.
In this study, we directly transfected synthetic siRNA oligos, which allowed us to evaluate their therapeutic effect on cancer cells. In fact, siRNAs can also be expressed from plasmid or viral vectors using the RNA polymerase III promoter (20, 22), and the expression may be combined with a tumor-specific promoter or an inducible system (23) such that siRNA can specifically target oncogenes in cancer cells without affecting normal cells. Future studies will investigate whether cyclin E overexpression can be efficiently depleted by siRNA expressed from a DNA-based expression vector.
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.
This work was supported by grants from the NIH (to P. S.), the American Cancer Society (to K. L.), and the Curtis Hankamer Basic Research Fund (to K. L.).
The abbreviations used are: RNAi, RNA interference; dsRNA, double-stranded RNA; nt, nucleotide; HCC, hepatocellular carcinoma; siRNA, small interfering RNA; TUNEL, terminal deoxynucleotidyl transferase-mediated nick end labeling; BrdU, bromodeoxyuridine.