Oncogenic and tumor suppressive roles of polo-like kinases in human hepatocellular carcinoma

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Polo-like kinase (PLK) proteins play critical roles in the control of cell cycle progression, either favoring or inhibiting cell proliferation, and in DNA damage response. Although either overexpression or down-regulation of PLK proteins occurs frequently in various cancer types, no comprehensive analysis on their function in human hepatocellular carcinoma (HCC) has been performed to date. In the present study, we define roles for PLK1, PLK2, PLK3, and PLK4 during hepatocarcinogenesis. Levels of PLK1, as assessed by means of real-time reverse-transcription PCR and western blot analysis, were progressively increased from nonneoplastic surrounding liver tissues to HCC, reaching the highest expression in tumors with poorer outcome (as defined by the length of patients' survival) compared with normal livers. In sharp contrast, PLK2, PLK3, and PLK4 messenger RNA and protein expression gradually declined from nontumorous liver to HCC, with the lowest levels being detected in HCC with shorter survival. In liver tumors, PLK2-4 down-regulation was paralleled by promoter hypermethylation and/or loss of heterozygosity at the PLK2-4 loci. Subsequent functional studies revealed that PLK1 inhibition led to suppression of cell growth in vitro, whereas opposite effects followed PLK2-4 silencing in HCC cell lines. In particular, suppression of PLK1 resulted in a block in the G2/M phase of the cell cycle and in massive apoptosis of HCC cells in vitro regardless of p53 status. Conclusion: PLK1-4 proteins are aberrantly regulated and possess different roles in human HCC, with PLK1 acting as an oncogene and PLK2-4 being presumably tumor suppressor genes. Thus, therapeutic approaches aimed at inactivating PLK1 and/or reactivating PLK2-4 might be highly useful in the treatment of human liver cancer. (HEPATOLOGY 2010.)

Polo-like kinase (PLKs) proteins play pivotal roles in cell cycle progression and response to DNA damage.1 Four members of this family of serine/threonine kinases were identified: PLK1, PLK2 (also known as SNK), PLK3 (also known as FNK or PRK), and PLK4 (or SAK).1 PLKs are characterized by a highly conserved N-terminal serine/threonine kinase domain and one or two polo boxes in the C-terminal region, which are crucial for subcellular localization and binding of specific phosphopeptides.2 Expression of PLKs is tightly regulated during the cell cycle.1 PLK1 is inhibited by numerous checkpoint genes, whereas PLK2-4 genes are activated by spindle checkpoints and DNA damage.1, 3 Despite the high sequence homology among the four members of the PLK family, their functions seem to diverge. PLK1 is involved mainly in the control of the G2/M phase, by promoting CDC25C phosphatase activity with subsequent activation of CyclinB1/CdK1 complex, and the degradation of early mitotic inhibitor-1 (EMI1), which inhibits the activated Anaphase-Promoting Complex/Cyclosome.1 PLK2 and PLK3 were identified as serum-inducible growth responsive genes and are implicated in the stress-response.4 Analysis of PLK2 knockout mice indicated that PLK2 is implicated in embryonic development and cell cycle regulation, as confirmed by recent findings showing an involvement of PLK2 in promotion of S-phase entry and centriole duplication.5 Previously, levels of PLK3 have been described as either unchanged throughout the cell cycle6 or increased in mitosis.7 However, more recent evidence indicates that PLK3 expression peaks in G1 phase and is required for S phase entry through the regulation of cyclin E levels.8 Moreover, PLK3 is implicated in the regulation of Golgi apparatus fragmentation during cell cycle progression, and deregulated expression of PLK3 in vitro promotes cell cycle arrest and apoptosis, mainly due to microtubule disfunctions.9, 10 Like all other PLKs, PLK4 is implicated in cell cycle regulation, because constitutive PLK4 expression leads to decreased cell growth and multinucleation in vitro.11, 12 Indeed, PLK4 is involved in the proper reproduction of centrosomes,13 and it is required for the APC-dependent destruction of cyclin B1, with the consequent exit from mitosis.11

Due to the critical role of PLKs in controlling cell cycle progression, their involvement in oncogenesis might be envisaged. An oncogenic role for PLK1 has been hypothesized, because its constitutive expression in NIH3T3 fibroblasts causes oncogenic foci formation and is tumorigenic in nude mice.14 Furthermore, PLK1 is overexpressed in a variety of human tumors,3 including human hepatocellular carcinoma (HCC).15, 16 Intriguingly, PLK1 expression directly correlates with tumor genetic instability and patient prognosis in several tumor types, suggesting that PLK1 up-regulation induces a mutator phenotype and contributes to tumor aggressiveness.3 In relation to their implication in carcinogenesis, less is known for PLK2, PLK3, and PKL4. A recent paper indicated that PLK2 is down-regulated by promoter hypermethylation in primary lymphomas and its overexpression in B cells lymphomas leads to apoptosis, suggesting that PLK2 act as a bona fide tumor suppressor gene.17 PLK3 expression has been also reported to diminish in some human tumors and it could contribute to generation of genetic instability, due to its role in the DNA damage response machinery.4 The antineoplastic function of PLK3 has been further substantiated by the observation that PLK3-deficient mice spontaneously develop tumors in various organs, including the liver.18 Recent evidence suggests a role for PLK4 as a tumor suppressor in hepatocarcinogenesis, becuase mice heterozygous for PLK4 (PLK4+/−) spontaneously develop liver and lung tumors.19 However, no comprehensive analysis on PLK proteins has been performed in human HCC to date. In this study, we investigated the status and the role of PLK proteins in a collection of human HCC as well as the molecular mechanisms responsible for modification of PLK levels in liver cancer. Our results indicate a deregulation of the four PLKs in human HCC, suggesting an oncogenic role for PLK1 and a tumor-suppressive function of PLK2, PLK3, and 4 in human hepatocarcinogenesis.

Abbreviations

FOXM1, forkhead box M1; HCC, hepatocellular carcinoma; HCCB, hepatocellular carcinoma with better outcome; HCCP, hepatocellular carcinoma with poorer survival; LOH, loss of heterozygosity; mRNA, messenger RNA; PLK, polo-like kinase; siRNA, small interfering RNA; SL, surrounding nontumorous liver.

Materials and Methods

Human Tissue Samples.

Six normal livers, 75 HCCs, and corresponding surrounding nontumorous liver tissues (SL) were used. Normal (disease-free) livers were from autopsy cases of healthy Caucasian individuals. Tumors were divided in HCC with shorter/poor survival (HCCP; n = 40) and longer/better survival (HCCB; n = 35), characterized by <3 and >3 years' survival following partial liver resection, respectively.20 Patient features are reported in Supporting Table 1. Liver tissues were kindly provided by Snorri S. Thorgeirsson (National Cancer Institute, Bethesda, MD). Institutional Review Board approval was obtained from participating hospitals and the National Institutes of Health.

Cell Lines and Treatments.

Human HCC cell lines (HepG2, HuH7, PLC, Hep3B, SNU-387, SNU-423, HLE, HuH6, SK-Hep1, and THLE-2), purchased from either the American Type Culture Collection or the Riken Cell Bank, were subjected to either small interfering RNA (siRNA) or demethylating treatments as reported in the Supporting Information.

Cell Viability, Apoptosis, and Cell Cycle Analyses.

Cell viability and apoptosis were determined using WST-1 Cell Proliferation Reagent and the Cell Death Detection Elisa Plus kit (Roche Molecular Biochemicals, Indianapolis, IN), respectively, according to the manufacturer's instructions. Analyses of cell cycle distribution were performed by means of flow cytometry of 2,4-diamidino-2-phenylindole–stained nuclei21 on a PAS II flow cytometer (Partec, Munster, Germany) using the Multicycle program (Phoenix Flow Systems, San Diego, CA). Experiments were repeated at least three times in triplicate.

Western Blot Analysis.

Hepatic tissue samples and cell lines were homogenized in lysis buffer and processed as described in the Supporting Information. Membranes were probed with specific primary antibodies (Supporting Table 2).

Quantitative Real-Time Reverse-Transcription Polymerase Chain Reaction.

Primers for PLK1-4 and ribonucleic acid ribosomal 18S (RNR-18, internal control) genes were obtained from Applied Biosystems (Foster City, CA). Polymerase chain reaction (PCR) and quantitative evaluations were performed as described in the Supporting Information.

Methylation-Specific PCR and Microsatellite Analysis.

Genomic DNA from normal livers, HCCs, and matching SL was modified using the EZ DNA methylation kit (Zymo Research, Orange, CA).22 Primers sets, used to assess the promoter status of PLK2 and PLK3 promoters, and PCR conditions were obtained from the literature.17 Primers to determine PLK4 methylation status were designed using the Methprimer software.23 Methylation-specific PCR and microsatellite analysis were performed as reported24 in the Supporting Information and Supporting Table 3.

Statistical Analysis.

Wilcoxon rank sum test and two-tailed Student t test were used to evaluate statistical significance. P < 0.05 was considered significant.

Results

PLK Family Members are Deregulated in Human HCC.

To assess the role of PLKs in human hepatocarcinogenesis, we first determined their levels in a collection of human normal livers, HCCs and respective nonneoplastic SL tissues using quantitative reverse-transcription PCR. A progressive up-regulation of PLK1 messenger RNA (mRNA) occurred from SL to HCC and was most pronounced in HCCs with shorter survival (HCCP) when compared with normal tissue (Fig. 1A). In sharp contrast, a significant decrease in PLK2 and PLK3 mRNA levels occurred in HCC when compared with corresponding SL, with the lowest levels of PLK2 and PLK3 being detected in HCCP (Fig. 1A). PLK4 expression gradually increased from SL to HCCB but was down-regulated in most HCCP (Fig. 1B). Results from western blot analysis closely resembled the data obtained by reverse-transcription PCR (Fig. 1B,C). No other clinicopathological features correlated with levels of PLK family members, including age, sex, etiology, presence of cirrhosis, tumor size, Edmondson/Steiner grade, and alpha-fetoprotein levels.

Figure 1.

Deregulation of the members of the PLK family and FOXM1 in human HCC. (A) Expression levels of PLK1-4 and FOXM1 in normal livers (NL, n = 6), HCC (35 HCCB and 40 HCCP, respectively), and corresponding nontumorous surrounding livers (SLB and SLP, respectively) by real-time reverse-transcription PCR. Surrounding livers from the two prognostic subclasses did not show statistical differences (P > 0.05). N-Target (NT) = 2 −ΔCt; ΔCt = RNR18-Ct target gene. (B) Representative Western blots of lysates prepared from normal livers, SL, HCCB, and HCCP and immunoblotted with the indicated antibodies. (C) Western blotting optical densities were normalized to β-actin values and expressed in arbitrary units. HCC and SL versus normal liver, at least P < 0.01; HCC subtypes versus corresponding surrounding liver, P < 0.001; *HCCP versus HCCB for P < 0.001 (Wilcoxon rank sum test).

Promoter Methylation and/or Loss of Heterozygosity are Linked to Down-regulation of PLK2, PLK3, and PLK4 in Human HCCs.

To investigate the molecular mechanisms responsible for down-regulation of PLK2, PLK3, and PLK4 in human HCC, we performed promoter methylation analysis for PLK2, PLK3, and PLK4 genes. No promoter methylation was detected for PLK2 and PLK3 genes in normal livers. In contrast, promoter methylation of PLK2 and PLK3 genes occurred in SL (Fig. 2A) and HCC (Fig. 2B,C). Methylation frequency for PLK2 was significantly higher in tumors (25/75 [33.3%]) than in nontumorous counterparts (3/75 [4%]; P = 8.55E-06). Additionally, HCCP showed a much higher frequency of PLK2 methylation than HCCB (18/40 [45%] versus 7/35 [20%]; P < 0.03) (Fig. 2B). PLK2 promoter methylation resulted in significantly reduced PLK mRNA (data not shown) and protein levels in HCC (22.3 ± 2.2 versus 48.0 ± 12.3; P = 2.03E-07).

Figure 2.

Frequency of promoter hypermethylation (HYP) and loss of heterozygosity (LOH) at PLK2, PLK3, and PLK4 genes in (A) human nontumorous surrounding livers and (B-D) corresponding HCC. Normal livers did not exhibit any altered profile and therefore are not shown. Presence of promoter hypermethylation and/or LOH or absence of the same alterations is indicated by black and gray squares, respectively. Both hypermethylated (B, C) and LOH positive cases (B-D) showed reduced expression of PLK2 (B), PLK3 (C), and PLK4 (D), respectively. Additionally, HCCB with PLK4 LOH showed higher PLK4 expression than LOH-positive HCCP. SLB and HCCB, nontumorous surrounding livers and corresponding HCC with better outcome, respectively; SLP and HCCP, nontumorous surrounding livers and corresponding HCC with poorer survival. *P = 0.001 (Wilcoxon rank sum test).

A similar trend was detected when assessing the frequency of PLK3 promoter methylation. Indeed, the PLK3 gene was silenced by promoter hypermethylation almost exclusively in HCC (28/75 [37.3%]) (Fig. 2C), whereas it was epigenetically inactivated by promoter hypermethylation in only two nonneoplastic surrounding livers (2.7%; P = 3.04E-08) (Fig. 2A). In HCC, frequency of promoter hypermethylation was significantly higher in HCCP (23/40 [57.5%]) than in HCCB (5/35 [14.3%]; P = 6.14E-05). Similar to PLK2, PLK3 levels were significantly reduced in HCCs with promoter hypermethylation (protein: 42.0 ± 7.8 versus 90.5 ± 13.3; P = 6.82E-10). In contrast, no PLK4 promoter hypermethylation was detected in any of the samples tested (Fig. 2D).

Genomic status of PLK2, PLK3, and PLK4 was further investigated through loss of heterozygosity (LOH) analysis of PLK2, PLK3, and PLK4 loci by comparing each HCC with respective SL. The LOH rates at PLK2, PLK3, and PLK4 gene loci were 20%, 24%, and 45.3%, respectively, and were always significantly more frequent in HCCP versus HCCB (Fig. 2B-D). Although LOH at the PLK2, PLK3, and PLK4 gene loci was statistically associated with reduced expression levels (15.9 ± 4.1 versus 45.4 ± 5.9 [P = 2.03E-07 for PLK2]; 42.2 ± 9.7 versus 81.8 ± 10.9 [P = 1.00E-05] for PLK3; 34.0 ± 6.3 versus 126 ± 18.6 [P = 3.82E-15] for PLK4, respectively), it showed a significant correlation with promoter hypermethylation for PLK2 (11/15 HCCs; Spearman's rho = 0.64; P = 8.45E-10) and PLK3 (13/19 HCCs; rho = 0.42; P = 1.84E-04), respectively, suggesting the inactivation of both alleles in these cases.

The role of methylation on PLK2 and PLK3 expression was further investigated in vitro. First, we screened 11 HCC cell lines for PLK2 and PLK3 promoter methylation. PLK2 methylation was detected in HepG2, HuH7, and Hep3B cell lines, whereas PLK3 methylation was detected in HepG2, HuH7, Hep3B, and SNU-387 cells (Supporting Fig. 1A). Subsequent treatment with the demethylating agent 5-AZA-cytidine caused a dose-dependent up-regulation of PLK2 and PLK3 mRNA in HepG2 and Hep3B (harboring PLK2 and PLK3 methylated promoter), but not in PLC (harboring PLK2 and PLK3 unmethylated promoter; Supporting Fig. 1B,C) cells.

Divergent Roles of PLKs on HCC Cell Growth.

The role of PLK family members in HCC cell growth was investigated by assessing the consequence of their inactivation by siRNA in HCC cell lines. Suppression of PLK1 expression resulted in a marked decrease of cell viability in HepG2 (p53 wild-type) and Hep3B (p53 deleted) cell lines (≈60% and 80%, respectively) when compared with untreated cells (Fig. 3A). This effect was associated with a marked induction of apoptosis in HepG2 (three-fold) and Hep3B cells (five-fold), respectively (Fig. 4A). An opposite effect was obtained when silencing the PLK2, PLK3, or PLK4 gene by siRNA in SNU-423 and HLE cell lines (expressing high levels of the PLK2, PLK3, and PLK4 genes). Indeed, suppression of PLK2, PLK3, or PLK4 was accompanied by a significant growth acceleration in the two cell lines (Fig. 3B-D) and resistance to apoptosis (Fig. 4B-D), suggesting that down-regulation of PLK2, PLK3, and PLK4 play a protumorigenic role in human hepatocarcinogenesis. Next, we assessed the possible interplay between PLKs by determining the levels of PLK1-4 genes following siRNA-mediated silencing of the other members of the PLK family. Interestingly, suppression of both PLK2 and PLK3 led to up-regulation of PLK1 (Supporting Figs. 2 and 3), implying a modulatory role of PLK2 and PLK3 over PLK1 expression. No additional modifications in gene expression were detected following silencing of PLK1 and PLK4 by siRNA (Supporting Figs. 2 and 3). Thus, the present data suggest that PLK1 promotes the growth of human HCC cells, whereas the down-regulation of PLK2, PLK3, and PLK4 antagonizes the antiproliferative and proapoptotic functions exerted by these proteins in nontumor cells.

Figure 3.

Effect of siRNA-mediated silencing of PLK1-4 genes on cell viability of human HCC cell lines. Cell viability was assessed using gene-specific and control siRNAs after 48 hours of treatment. PLK1 was silenced in HepG2 and Hep3B cell lines (A), whereas PLK2-4 were suppressed in SNU-423 and HLE cells (B-D). Untreated cells were used for normalization. Each bar represents the mean ± standard deviation (n = 9). siPLK1-treated cells versus transf C, at least P < 0.001; siPLK1-treated cells versus nonsense, P < 0.001 (Student t test). C, untreated cells; ns, nonsense siRNA or control siRNA; siPLK1-4, siRNA against PLK1-4; transf C, transfection reagent control.

Figure 4.

Effect of siRNA-mediated silencing of PLK1-4 genes on apoptosis in human HCC cell lines. Apoptosis was assessed using gene-specific and control siRNAs after 48 hours of treatment. PLK1 was silenced in HepG2 and Hep3B cell lines (A), whereas PLK2-4 were suppressed in SNU-423 and HLE cells (B-D). Untreated cells were used for normalization. Each bar represents the mean ± standard deviation (n = 9). siPLK1-treated cells versus transf C, at least P < 0.001; siPLK1-treated cells versus ns, P < 0.001 (Student t test). Abbreviations: C, untreated cells; ns, nonsense siRNA or control siRNA; siPLK1-4, siRNA against PLK1-4; transf C, transfection reagent control.

PLK1 Inhibition Induces G2/M Phase Block and Apoptosis in HCC Cell Lines.

Because the most pronounced antitumorigenic effects on HCC cell growth were obtained by targeting PLK1, our following studies focused on the role of PLK1 in the regulation of cell cycle and apoptosis in HCC cells. Silencing of PLK1 expression by siRNA in Hep3B and HepG2 cells resulted in a block in G2/M phase (Fig. 5A) as well as in a strong increase of the sub-G1 fraction indicating apoptosis (data not shown), as confirmed by the detection of cleaved PARP protein (Fig. 5B). In addition, inhibition of PLK1 expression was followed by down-regulation of the antiapoptotic protein survivin (Fig. 5B), supporting the recent finding that PLK1 promotes cell survival through inhibition of survivin degradation in esophageal cancer cells.25 Previous evidence indicated that PLK1 can bind to p53 and abrogate its tumor suppressor functions,26 and recent reports have demonstrated that PLK1 is able to phosphorylate the tumor suppressor p73, with consequent inhibition of its transcriptional activity, thereby suppressing apoptosis.27, 28 Thus, we determined whether the activation of p53 and p73 proteins could be involved in the apoptotic response following PLK1 inhibition. In accordance with our hypothesis, up-regulation of p53 and p73 protein levels as well as activation of their target genes p21CIP1 and BAX was detected in HepG2 cells (p53 wild-type) following PLK1 inhibition (Fig. 6A). In Hep3B cells (p53 deletion), apoptosis induction was paralleled by a rise in p73 expression and the induction of p21CIP1 and BAX (Fig. 6A). Furthermore, siRNA-mediated silencing resulted in BAX activation in HepG2 and Hep3B cells, as demonstrated by its translocation to the mitochondria and subsequent release of cytochrome C into the cytoplasm (Fig. 6B). The involvement of an intrinsic apoptosis pathway following PLK1 depletion in both cell lines was also confirmed by a decrease of MCL1 and BCL2 antiapoptotic proteins, as well as an increase of active (cleaved) caspase-3 (Fig. 6C). Taken together, these findings support a role of PLK1 in HCC progression by its ability to antagonize apoptosis dependent on p53 family members.

Figure 5.

Effect of siRNA-mediated silencing of PLK1 on cell cycle and apoptotic cascades in human HCC cell lines. Western blot analysis revealed that inhibition of PLK1 expression results in a G2/M block (A) as well as increased PARP cleavage and a decline of the survivin levels (B), indicating apoptosis induction, when compared with control siRNA-treated cells. Abbreviations: C, untreated cells; ns, nonsense siRNA or control siRNA; PARP cl, cleaved PARP; PARP fl, full-length PARP; siPLK1, siRNA against PLK1; transf C, transfection reagent control.

Figure 6.

Mediators of apoptosis following PLK1 in human HCC cell lines either wild-type (HepG2) or deleted (Hep3B) for the p53 gene. (A) Induction of p53 and p73 genes and their putative target proteins, p21 and BAX, were detected by means of Western blotting in cells transfected with siRNA against PLK1 for 48 hours. An involvement of the mitochondrial apoptotic pathway was confirmed by way of western blotting analysis of (B) BAX and cytochrome C (Cytochr C) localization, and of (C) MCL1, BCL2, and cleaved caspase-3 protein levels (CASPASE cl). Abbreviations: C, untreated cells; ns, nonsense siRNA or control siRNA; siPLK1, siRNA against PLK1; transf C, transfection reagent. The arrows indicate the cleaved CASPASE3 products.

PLK1 Expression Is Regulated by Ha-Ras Through Forkhead Box M1 in HCC Cells.

Previously, it has been demonstrated that PLK1 is a target gene of the forkhead box M1 (FOXM1) transcription factor in the mouse liver.29 Thus, we investigated FOXM1 at mRNA and protein levels in our HCC samples (Fig. 1A-C). From these analyses, a progressive increase of FOXM1 levels with a similar trend to that observed for PLK1 was detected from SL to HCC, which is in accordance with a previous report.30 Furthermore, in order to explore whether FOXM1 can influence PLK1 expression in human HCC, we assessed the consequence of modulating FOXM1 expression on PLK1 protein levels in vitro. Overexpression of FOXM1 in the SNU-182 HCC cell line (exhibiting low FOXM1 mRNA levels) led to up-regulation of PLK1 protein (Fig. 7A). Conversely, PLK1 down-regulation occurred when FOXM1 expression was inhibited by specific siRNA in Hep3B and HepG2 cells (displaying high FOXM1 gene expression) (Fig. 7B). A recent report indicated that PLK1 might be indispensable for the growth of K-Ras–mutated cells in various tumor types.31 Therefore, we determined whether PLK1 is necessary for Ras-induced cell growth of HCC in vitro. First, we overexpressed both the wild-type and the mutated form (substitution of leucine for glutamine at position 61, Q61L) of Ha-Ras, the most significantly up-regulated gene of the Ras family in HCC,32 in SNU-182 cells by means of transient transfection. Forced induction of both wild-type and mutant Ha-Ras led to increased levels of FOXM1 and PLK1 in SNU-182 cells (Fig. 7C), suggesting that PLK1 lies downstream of a cascade initiated by Ha-Ras and propagated by FOXM1 in HCC. In accordance with the latter hypothesis, the induction of PLK1 by wild-type and mutant Ha-Ras was suppressed when transient transfection of Ha-Ras was coupled to FOXM1 inhibition by siRNA (Fig. 7C). As a consequence, a strong suppression of SNU-182 in vitro growth was detected when transfection of either wild-type or mutated Ha-Ras was paralleled by suppression of FOXM1 expression by siRNA (Fig. 7D). A similar, remarkable constraint of Ha-Ras–induced cell growth was detected when transfection of either wild-type or mutated Ha-Ras was coupled to PLK1 silencing by siRNA (Fig. 5D). Together, these data indicate that PLK1 up-regulation is driven by a Ha-Ras/FOXM1 cascade and underline the importance of an intact Ha-Ras/FOXM1/PLK1 axis in sustaining the growth of human HCC cell lines.

Figure 7.

Effect of FOXM1 modulation on PLK1 protein levels in human HCC cell lines. (A) Overexpression of FOXM1 by transient transfection in SNU-182 cells up-regulates PLK1 levels. (B) Silencing of FOXM1 in Hep3B cells determines down-regulation of PLK1 levels. Equivalent results were obtained in HepG2 cells (data not shown). (C) Overexpression of wild-type or oncogenic mutant (Q61L) Ha-Ras leads to up-regulation of PLK1 through FOXM1. The SNU-182 cell line was transfected with either wild-type Ha-Ras or its mutated form (Q61L) alone, or associated with siRNA against FOXM1 (siFOXM1) or PLK1 (siPLK1). Representative Western blotting panels at 48 hours after transfection are shown. (D) Effect of transfection with wild-type Ha-Ras, its mutated form (H-Ras Q61L) alone, or associated with siRNA against FOXM1 (siFOXM1) or PLK1 (siPLK1), on in vitro growth of SNU-182 cells. Results at 48 hours after transfection are expressed as the mean ± standard deviation. Untreated cells (Control) were used for normalization. Experiments were performed at least three times in triplicate. *Wild-type Ha-Ras associated with siFOXM1 or siPLK1 versus Ha-Ras alone, P < 0.001; mutated Ha-Ras Q61L associated with siFOXM1 or siPLK1 versus Ha-Ras Q61L alone, P < 0.001 (Student t test).

Discussion

Human HCC is one of the most frequent and lethal tumors worldwide.33 Despite new therapeutic strategies,34 the life expectancy of patients with unresectable HCC remains poor. Thus, the identification of new molecular therapeutic targets is mandatory to improve patient outcome.

Here, we show for the first time that members of the PLK family—namely PLK1, PLK2, PLK3, and PLK4—are aberrantly expressed in human HCC. A gradual up-regulation of PLK1 expression was observed from normal liver to HCC, in accordance with previous reports.15, 16 Moreover, a significant increase of PLK1 expression occurred in HCCP when compared with HCCB, implying a role of PLK1 in HCC biological aggressiveness and patient outcome. In accordance with the latter hypothesis, it has been recently demonstrated that the PLK1-Cdc25A pathway is aberrantly activated in human HCC and enhances the metastatic potential of liver tumors.15

Interestingly, our data show an opposite trend of expression for PLK2, PLK3, and PLK4 genes at mRNA and protein levels than that of PLK1 in HCC. Indeed, the lowest levels of PLK2, PLK3, and PLK4 were detected in the more aggressive HCCs. This finding, together with the divergent effects on HCC cell growth, supports the hypothesis that, despite belonging to the same family of kinases, they might play strikingly opposite roles in oncogenesis. Indeed, PLK1 inactivation led to decreased cell viability and a rise in apoptosis in HCC cell lines, whereas an increase in cell growth and a decline in apoptosis followed the silencing of PLK2, PLK3, and PLK4 genes.

Subsequent studies on PLK1, PLK2, PLK3, and PLK4 genes indicate that different mechanisms can influence the levels of these genes in human liver cancer. We demonstrated that PLK1 expression is driven by the protooncogene Ha-Ras via the FOXM1 transcription factor in human HCC cell lines. The dependence of PLK1 from FOXM1 activity has been observed in a mouse liver model.29 Furthermore, because FOXM1 and PLK1 regulate each other's activity,35 this finding might explain the concomitant up-regulation of FOXM1 and PLK1 detected in human HCC samples. In addition, these data substantiate our previous observations assigning a key role to FOXM1 in HCC growth and prognosis, due to its ability to modulate a large subset of genes involved in cell cycle progression.30 Of note, we found that PLK2 and PLK3 can negatively regulate PLK1 expression in HCC cell lines, suggesting the existence of a control mechanism on PLK1 levels that must be subverted to achieve unrestrained PLK1 expression in liver cancer. Additional studies are required to better define the mechanisms exerted by PLK2 and PLK3 to down-regulate PLK1.

Negative regulation of PLK2 and PLK3 genes seems to depend mainly on promoter hypermethylation in human HCC, especially in HCCP, and LOH may represent the second hit for complete PLK2 inactivation. Considering that a similar regulation of PLK2 was found in human lymphomas,17 our data support the idea that PLK2 hypermethylation and LOH at the PLK2 locus might be major complementary mechanisms responsible for the inactivation of PLK2 in cancer. PLK3's inactivation through promoter hypermethylation and/or LOH is similar to that described for PLK2 and substantiates a possible role of PLK3 as a liver tumor suppressor gene. Accordingly, silencing of the PLK3 gene triggered hepatocarcinogenesis in a mouse model.18 Moreover, we frequently found LOH for the PLK4 gene in many HCC samples, with the highest incidence in HCCP. The PLK4 locus is located at the chromosomal band 4q28.1, which is frequently affected by LOH in HCC and whose crucial role in liver carcinogenesis has been envisaged.36, 37 In accordance with the latter hypothesis, PLK4 heterozygosity resulted in spontaneous liver tumor development in a mouse model, which was associated with centrosome amplification and induction of chromosomal instability19 as characteristically observed in human HCC.37, 38 Thus, PLK4 might be one of the pivotal tumor suppressor genes located in the 4q28.1 chromosome region, whose loss contributes to human hepatocarcinogenesis.

Furthermore, we have investigated in more detail the role of PLK1 on cell cycle regulation in human HCC cell lines. Our data confirm the important function of PLK1 in regulating both the G2/M phase of the cell cycle and the apoptotic process, supporting previous observations in various cancer cell lines.25, 39, 40 In particular, the present findings indicate that PLK1 is able to inhibit apoptosis in a p53 family–dependent manner, as observed in Hep3B and HepG2 cell lines. It has been demonstrated that PLK1 interacts with the DNA binding domain of p53, thereby decreasing its stability and transcriptional activity.26 The latter mechanism might explain the increased apoptosis rate reported in HepG2 cells (p53 wild-type) with subsequent down-regulation of antiapoptotic proteins following PLK1 silencing. Recently, a physical interaction between PLK1 and p73, another member of the p53 family, has been demonstrated in different cell lines.27, 28 Like p53, p73 transactivates many p53 target genes involved in cell cycle control and apoptosis. PLK1 is able to phosphorylate p73 at the threonine 27 residue within its transactivation domain, thereby abrogating its transcriptional activity.27, 28 We detected an increase in p73 protein level and its target genes following silencing of PLK1 expression in Hep3B and HepG2 cells. Up-regulation of the p73 protein was also observed in MCF7 breast cancer cells expressing the p53 gene,27 confirming that p73 induction by PLK1 is independent of p53 in different cellular contexts.

In a recent report, a therapeutic approach using a PLK1 inhibitor resulted in dramatic tumor regression in nude mice bearing xenografts of HCT116 colorectal cancer cells in which the p53 gene was disrupted, suggesting a crucial function of PLK1 for the growth of p53-deficient tumor cells.41 Similarly, we show here that the growth of SNU-182 cells overexpressing Ha-Ras, FOXM1, and PLK1 is dramatically reduced and impaired when this axis is disrupted by either FOXM1 or PLK1 suppression through siRNA in vitro (Fig. 7D). Therefore, PLK1 might represent a fundamental downstream effector of several oncogenic pathways in HCC, as described in colorectal cancer cell lines.31 Furthermore, in the HCT116 xenograft cancer model, suppression of PLK1 resulted in a striking reduction of in vitro growth of cell lines harboring K-Ras mutations, but not in wild-type K-Ras cells.31 In SNU-182 cells instead, we found that suppression of either PLK1 or its upstream inducer FOXM1 strongly suppresses the growth of Ha-Ras overexpressing cells regardless of Ha-Ras mutation status. The latter finding supports a crucial, indispensable growth-promoting stimulus by PLK1 in oncogenic cascades activated by wild-type Ras as well. In human HCC, mutations of the Ras genes are extremely rare, but multiple mechanisms other than somatic mutations lead to unconstrained Ras activity.32 Therefore, PLK1 might be a crucial therapeutic target in HCC, due to the ubiquitous activation of the Ras pathway in this disease.32

In conclusion, our data clearly demonstrate that PLK1 plays oncogenic functions, whereas PLK2-4 are presumably tumor suppressor genes in human hepatocarcinogenesis. Combination of PLK1 up-regulation and PLK2-4 down-regulation may have a central role in unrestrained cell cycle progression and, consequently, in proliferation of human HCC cells. Thus, therapeutic approaches aimed at suppressing PLK1 and/or reactivating PLK2-4 genes might be highly beneficial for the treatment of human HCC.

Acknowledgements

We thank Snorri S. Thorgeirsson (National Cancer Institute, Bethesda, MD) for providing human liver tissue samples.

Ancillary