Polo-like kinase 1 (Plk1) is required for multiple stages of mitosis and is up-regulated in many human malignancies. We depleted Plk1 expression using small interfering RNA (siRNA) and showed defects in bipolar spindle formation and cytokinesis, growth inhibition, and apoptosis induction in human cancer cell lines. To our surprise, depletion of Plk1 in normal human cells did not result in obvious cell cycle defects, and did not induce significant inhibition of cell growth for at least two cell cycles. In addition, Plk1 siRNA inhibited colony formation in soft agar and tumorigenesis in a HT1080 xenograft model in a dose-dependent manner. Analysis with isogenic pairs of cell lines, differing in p53 status, revealed that Plk1 depletion preferentially induced mitotic arrest, aneuploidy, and reduced cell survival in the p53-defective cell lines. No obvious defects were observed in most p53 wild-type cells during the first few cell cycles. In addition, long-term survival studies revealed that p53 facilitates survival upon Plk1 depletion. Therefore, short-term inhibition of Plk1 can kill tumor cells while allowing normal cells to survive. These data validate the episodic inhibition of Plk1 as a very useful approach for cancer treatment.
Polo-like kinase 1 (Plk1) is a serine/threonine kinase that functions to regulate many stages of mitosis (1). Activation of the cyclin B/Cdc2 complex triggers entry into mitosis. Plk1 phosphorylates and activates the Cdc25C phosphatase (2–4) , which reverses the inhibitory phosphorylation of Cdc2. Additionally, Plk1 phosphorylates cyclin B during prophase (5, 6) and, later in mitosis, facilitates its degradation and progression into cytokinesis. Plk1 is required for centrosome maturation, bipolar spindle formation, and chromosomal segregation, and is associated with microtubules and centrosomes at various stages of mitosis (1, 7–10) . Microinjection of Plk1-neutralizing antibody results in monoastral arrays, leading to chromosome missegregation (8). Plk1 also regulates cytokinesis (11–14) and has been shown to localize to the midbody of the cleavage furrow (15). Furthermore, Plk1 responds to DNA damage checkpoints (16–21) .
Plk1 is overexpressed in a variety of human malignancies (22–26) , and its expression levels are often a prognostic factor for disease progression (27–30) . In addition, the overexpression of Plk1 in NIH3T3 cells can transform cells and induce tumors in nude mice (31). Blocking Plk1 expression or function has been shown to induce cytotoxicity in cancer cells (32–37) . However, potential toxicity in normal tissues poses a significant risk for Plk1 inhibitors as cancer therapeutics. Overexpression of Plk1 in human malignancies indicates that tumor cells may depend on higher Plk1 levels for their growth and survival. Therefore, it is conceivable that Plk1 inhibition could confer specific cytotoxicity to tumor cells. By depleting Plk1 with small interfering RNA (siRNA) in human tumor cells, we have shown a critical role for Plk1 in tumor cell survival. Plk1 siRNA also dramatically delayed tumor growth in severe combined immunodeficient (SCID) mice. In contrast, normal human cells survived Plk1 depletion very well without apparent defects during the first few cycles. In addition, we identified p53 as one factor contributing to the differential cytotoxicity. These data provide critical evidence that episodic Plk1 inhibition could confer specific cytotoxicity toward tumor cells.
Materials and Methods
Cell lines. All cell lines, except 184B5-E6, were purchased from the American Type Culture Collection (Manassas, VA). 184B5-E6 was a generous gift from Dr. J. Gudas (National Cancer Institute).
Small interfering RNA sequences. Plk1 siRNAs were purchased from Dharmacon Research, Inc. (Lafayette, CO): Plk1-Si1 (AACCAGUGGUUCGAGAGACAG) and Plk1 Si1C (AACCAGUUUGGCGAGAGACAG). The selection of these sequences was based on the sequence of the best Plk1 AS oligo isis121969 (GAACCAGUGGUUCGAGAGACA) obtained from ISIS Pharmaceuticals (Carlsbad, CA; data not shown).
Transfection. siRNA was transfected using LipofectAMINE 2000 (LF2000) from Invitrogen (Carlsbad, CA) according to the manufacturer's instructions.
Western blotting. Transfected cells were harvested and lysed in insect cell lysis buffer [10 mmol/L Tris (pH 7.5), 130 mmol/L NaCl, 1% Triton X-100, 10 mmol/L NaF, 10 mmol/L NaPi, 10 mmol/L NaPPi] supplemented with 50× protease inhibitor cocktail (BD PharMingen, Bedford, MA) and 1 μmol/L microcystin LR (Sigma Chemical Co., St. Louis, MO). Immunoblotting was done with anti-Plk1 antibody from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). A nonspecific band recognized by Plk1 antibody or actin was used as the loading control.
Bipolar spindle morphology assay. Cells were washed with PBS and fixed with 50% methanol/50% acetone for 2 minutes at room temperature. After washing with PBS, the cells were blocked with PBS containing 3% bovine serum albumin at room temperature for 20 minutes. The cells were then stained with FITC-labeled anti-α-tubulin antibody (Sigma) and mounted on slides with Cytofluor containing 4′,6-diamidino-2-phenylindole (Ted Pella, Inc., Redding, CA).
AlamarBlue assay. AlamarBlue assays were carried out as per the manufacturer's instructions (Biosource International, Inc., Camarillo, CA). Analysis was done using a fmax Fluorescence Microplate Reader (Molecular Devices, Sunnyvale, CA), set at the excitation wavelength of 544 nm and emission wavelength of 595 nm. Data was analyzed using SOFTmax PRO software. Each data point is the average of six values. Error bars represent the SD.
Statistical analysis. Pairwise, two-tailed, unequal variance Student's t test was done to assess the difference between the effect of Si1 and Si1C. Asterisks indicate a significant difference with P < 0.01 in 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide and P < 0.05 in clonogenic assay. P values are listed in Supplementary Fig. 4.
Xenograft model. HT1080 human fibrosarcoma cells were transfected with Plk1 siRNA Si1 or Si1C. At 24 hours post-transfection, live cells were harvested and mixed with an equal volume of Matrigel (Collaborative Biomedical Products, Bedford, MA). Cells (0.5 million) were inoculated into the flank of SCID-Beige mice (Charles River Laboratories, Wilmington, MA, USA). Tumor volume was measured twice weekly with digital calipers. Tumor volume was estimated using the formula: V = L × W2 / 2. Each treatment group contained 12 mice.
Caspase assay. Caspase assays were carried out as previously described (38). Each data point is the average of three values. Error bars represent the SD.
Clonogenic assay. Cells were plated at 1,000 cells per dish at 24 hours post-transfection and incubated for 10 to 14 days at 37°C. Colonies were quantified with Image-Pro Plus (Media Cybernetics, Silver Spring, MD) after fixing with 100% methanol and stained with 5% Giemsa stain (Sigma). Each data point is the average of two values. Error bars represent the range of values.
Flow cytometry analysis. Attached or detached cells were harvested and pelleted by centrifugation at 800 × g for 5 minutes at 4°C. The cells were washed with PBS and resuspended in 0.5 mL ice-cold staining solution (5 μg/mL propidium iodide, 40 units/mL RNase A, 0.25% Triton X-100, in PBS). After a 1-hour incubation at 4°C in the dark, the DNA content was analyzed using a Becton Dickinson ExCalibur Flow Cytometer (San Jose, CA).
Plk1 depletion induces growth inhibition and apoptosis in human tumor cells but has no significant effect on normal cells during the first two cell cycles. Human large cell lung cancer cell line H1299 was transfected with Plk1 siRNA Si1 or its control Si1C. Plk1 expression was completely blocked by Si1, whereas Si1C had no effect ( Fig. 1A ). Additionally, Si1 reduced overall cell survival and induced caspase activation when compared with Si1C ( Fig. 1B and C). Because Si1C alone caused small levels of nonspecific cytotoxicity ( Figs. 1 - 6 ), we assessed the effects of Plk1 depletion by comparing the differences between Si1 and Si1C. Cellular phenotypes occurred within the same concentration range of Si1 required to block Plk1 expression. Forty-eight hours after transfection, Si1 transfectants were arrested in G2-M and displayed monoastral spindle arrays ( Fig. 1D and E). Approximately 90% of Si1-treated mitotic cells exhibited such spindle defects. Si1C transfectants or reagent control-treated cells showed normal spindle morphology and cell cycle distribution. A significant increase in multinuclear cells was also observed in Si1 transfectants ( Fig. 1F).
We extended our studies to additional human tumor cell lines representing different tissue types. These included the pancreatic carcinoma line MiaPaCa and the cervical carcinoma line HeLa. In all cases, Plk1 depletion resulted in G2-M arrest and a dramatic reduction in cell survival ( Fig. 2). We also examined the effect of Plk1 siRNA on normal human cells, including normal human fibroblasts (NHF), foreskin fibroblasts (IMR90), and mammary gland epithelial cells (184B5, see below). In all cases, Plk1 siRNA blocked Plk1 expression within a similar concentration range ( Fig. 2A). However, the impact on cell survival and cell cycle distribution was minimal for the first two cycles of cell division ( Fig. 2B and C). The effect of Plk1 depletion on long-term cell growth was examined in NHF and IMR90 cells. In cells cultured for 96 hours, which is long enough for three division cycles for NHF and two division cycles for IMR90, Plk1 depletion caused a minor increase in G2-M accumulation and growth inhibition ( Fig. 3A and B). However, the efficiency of NHF colony formation was dramatically reduced by Si1, indicating that the long-term viability of these cells requires Plk1 function ( Fig. 3C). These results indicate that normal cells could survive for at least two division cycles without Plk1 function, whereas cancer cells depend on Plk1 in every cycle. This differential impact of Plk1 depletion on cell survival may be due to its differential effect on G2-M arrest in the two sets of cells. We conclude that episodic inhibition of Plk1 may confer cancer-specific cytotoxicity.
Plk1 depletion preferentially confers cytotoxicity toward p53-defective cells. Virtually all cancer cells are defective in either the p53 or Rb pathways (39, 40) . To explore whether p53 contributes to the differential sensitivity observed upon Plk1 depletion, we compared the sensitivity of isogenic cell lines differing in p53 status. Human mammary gland epithelial cells (184B5), colon cancer cells (RKO), and their E6-transformed counterparts were chosen for this study (41). Supplementary Fig. 1A shows the presence of wild-type p53 function in 184B5 cells. p53 and p21 levels were induced by doxorubicin in 184B5 cells, but not in the p53-defective 184B5-E6 cells. We observed a significant difference in cell survival between 184B5 and 184B5-E6 upon the same level of Plk1 depletion ( Fig. 4A and B). Similar results were obtained in the RKO and RKO-E6 pair (Supplementary Fig. 2A and B).
Fluorescence-activated cell sorting analysis revealed that Plk1 Si1 caused G2-M accumulation in 184B5-E6 and RKO-E6 cells, whereas Si1C had minimal effect. Additionally, there were significant amounts of polyploid cells in 184B5-E6 and RKO-E6 cells, suggesting that some of the E6-transformed cells proceeded to the next cell cycle even after the failure of cytokinesis ( Fig. 4C and Supplementary Fig. 2C). In contrast, G2-M accumulation in p53 wild-type cells was minimal upon Plk1 depletion at 48 hours post-transfection ( Fig. 4C and Supplementary Fig. 2C) and 72 hours post-transfection (data not shown). Si1 can maintain Plk1 depletion for at least 7 days post-transfection ( Fig. 4A). Both 184B5 and 184B5-E6 cells were actively growing for at least 4 days after transfections ( Fig. 5A). We also examined the impact of Plk1 depletion on the long-term survival of these cells. More 184B5 cells survived than 184B5-E6 cells in a clonogenic assay, suggesting that p53 wild-type cells survive much better than p53-defective cells upon Plk1 depletion ( Fig. 5B). As expected, some increase in G2-M was the result of mitotic arrest, as indicated by the increase in H3 phosphorylation (Supplementary Fig. 1B). In order to examine the possibility that the differential effects we observed were due to other effects of E6, we showed similar differential phenotypes between 293 and 293 T cells, in which p53 and RB functions are impeded by SV40 large T antigen (Supplementary Fig. 2D and E), supporting the hypothesis that p53 is one of the critical factors in the differential response.
The functions of p53 in cell cycle arrest and apoptosis induction were examined in the 184B5 pair treated with Plk1 siRNA. p53 levels were not changed in the parental 184B5 cells, whereas a very minor but reproducible increase was observed in the 184B5-E6 cells when treated with 10 nmol/L Si1. This is consistent with a previous report on Plk1 siRNA (32). However, neither p21 nor Bax was induced in either cell line ( Fig. 4D), and no increase in G1 phase was detected ( Fig. 4C).
Plk1 depletion by small interfering RNA reduces tumor growth in a xenograft model. We chose the HT1080 (human sarcoma cell line) xenograft model to examine the effect of Plk1 depletion on tumor growth. Depletion of Plk1 by Si1 resulted in a dramatic, dose-dependent reduction in HT1080 survival and colony formation ( Fig. 6A-C). While assessing the duration of Si1-mediated Plk1 depletion, we observed that most of the cells die 3 to 4 days after transfection, whereas Plk1 is absent. However, after 5 days, cells grew out from sparse cell islands and Plk1 levels returned to normal (data not shown). We suspect that these islands represent untransfected cells.
The effect of Plk1 depletion on tumor growth was then assessed in vivo. Tumor growth from Si1-transfected HT1080 cells was significantly delayed in a dose-dependent manner when compared with Si1C-treated samples. In fact, there was little growth within 14 or 17 days in tumors treated with 10 or 30 nmol/L Si1, respectively. After 19 days, the growth rates of tumors treated with Si1 was indistinguishable from the controls ( Fig. 6D). We suspect that Plk1 levels in these tumors return to control levels due to Si1 loss or the outgrowth of untransfected cells. The earliest we could harvest tumors big enough to measure Plk1 levels was 5 days post-inoculation, at which time Plk1 mRNA levels in Si1-tumors were the same as controls (data not shown). At the end of the study, Plk1 protein levels in Si1-tumors were the same as in the controls ( Fig. 6E). Because we inoculated the same amount of live cells from each treatment group, cells transfected with Si1 probably died during the first few days, as suggested by the experiments in vitro. The tumors that eventually grew out from the Si1-treated group were likely derived from untransfected cells. These data show the requirement of Plk1 function in maintaining tumor growth.
We have shown that Plk1 siRNA disrupts bipolar spindle formation, increases the number of multinuclear cells, and effectively reduces cell survival in human cancer cell lines. These results are consistent with previous reports on Plk1 siRNA, antisense oligos, peptide inhibitors, and a Plk1 dominant-negative mutant (32–37) . In addition, we have shown that Plk1 siRNA inhibits tumor formation by HT1080 cells in SCID mice, suggesting that Plk1 is required for tumor growth.
A major concern in targeting Plk1 for cancer treatment is the potential toxicity in normal tissues due to its pleiotropic functions in mitosis. To our surprise, Plk1 siRNA did not reduce cell survival in normal human cell lines, whereas it preferentially induced mitotic arrest and decreased cell survival in human tumor cell lines. In fact, normal cells were able to bypass the depletion of Plk1 and maintain active cycling without obvious cell cycle defects for at least two cell cycles ( Figs. 2 and 4). These results indicate that episodic Plk1 inhibition can specifically kill cancer cells within a reasonable therapeutic window.
p53 status may contribute to the differential sensitivity of normal and cancer cells to Plk1 depletion. Because these differences might be due to other genetic factors, we examined three sets of isogenic cell lines in which p53 and/or Rb function is blocked by either E6 or T antigen. Upon Plk1 depletion, the p53-defective cells showed increased G2-M arrest and decreased survival when compared to their p53 wild-type counterparts, both in short-term and long-term experiments ( Figs. 2 -5). However, transiently blocking p53 expression with siRNA was not sufficient to significantly increase the G2-M arrest and reduce cell survival upon Plk1 depletion (data not shown). It is possible that a p53 depletion–related event is modulating the response to Plk1 depletion, or that other genetic changes contribute to such modulation. For example, the absence of p53 function may allow other mutations to accumulate over time, which could modulate the response. It is also possible that transformation by E6 or T antigen results in the more profound G2-M arrest and decreased survival upon Plk1 depletion. Indeed, p53 is apparently not the only factor in this response, since cells such as 293 show significant G2-M arrest upon Plk1 depletion in the presence of functional p53. However, p53-defective/T antigen transformed 293 T cells still exhibit a more profound G2-M arrest and less survival when compared to 293 cells (Supplementary Fig. 2E).
Liu and Erikson (32) have reported that Plk1 depletion induces p53-dependent apoptosis in cancer cells. Our results were consistent with theirs since they reported p53 induction and apoptosis 3 days or later post-transfection in p53 wild-type cells. To take these studies further, we examined the kinetics of these responses and showed that p53 function or other factors may help normal cells to cope with the loss of Plk1 (through a yet unknown mechanism) for the first few days without apparent defects. Eventually, Plk1 depletion will induce cytotoxicity in these p53 wild-type cells, although much later than in p53-defective, oncogenic transformed cells. This may provide a reasonable therapeutic window for Plk1 inhibitors in cancer intervention.
As recently implied by van Vugt et al. (42), Plk1 depletion may invoke checkpoint responses due to defects in chromosomal condensation/segregation. Our preliminary results also indicate checkpoint responses upon Plk1 depletion. 3 p53 could conceivably affect cell survival via modulating such checkpoint responses.
In addition to p53, other mitotic kinases could compensate for the loss of Plk1 in normal cells. This compensation may not occur in p53-defective tumor cells or may simply not be sufficient. These tumor cells may require higher Plk1 activity, as evidenced by the frequent overexpression of Plk1 in multiple tumor types. Further experiments are required to test this hypothesis.
While this manuscript was under revision, Spänkuch et al. reported similar results, wherein Plk1 depletion reduced tumor growth in nude mice (43).
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 Dr. Paul Jung and Sandra-Xintong Hu for the QPCR study, Loren Lasko for the excellent technical support, and Dr. Yu Shen for providing siRNA.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
↵3 Unpublished results.
- Received June 16, 2004.
- Revision received December 27, 2004.
- Accepted January 24, 2005.
- ©2005 American Association for Cancer Research.