Aberrant methylation of Polo-like kinase CpG islands in Plk4 heterozygous mice
© Ward et al; licensee BioMed Central Ltd. 2011
Received: 11 March 2010
Accepted: 15 February 2011
Published: 15 February 2011
Skip to main content
© Ward et al; licensee BioMed Central Ltd. 2011
Received: 11 March 2010
Accepted: 15 February 2011
Published: 15 February 2011
Hepatocellular carcinoma (HCC), one of the most common cancers world-wide occurs twice as often in men compared to women. Predisposing conditions such as alcoholism, chronic viral hepatitis, aflatoxin B1 ingestion, and cirrhosis all contribute to the development of HCC.
We used a combination of methylation specific PCR and bisulfite sequencing, qReal-Time PCR (qPCR), and Western blot analysis to examine epigenetic changes for the Polo-like kinases (Plks) during the development of hepatocellular carcinoma (HCC) in Plk4 heterozygous mice and murine embryonic fibroblasts (MEFs).
Here we report that the promoter methylation of Plk4 CpG islands increases with age, was more prevalent in males and that Plk4 epigenetic modification and subsequent downregulation of expression was associated with the development of HCC in Plk4 mutant mice. Interestingly, the opposite occurs with another Plk family member, Plk1 which was typically hypermethylated in normal liver tissue but became hypomethylated and upregulated in liver tumours. Furthermore, upon alcohol exposure murine embryonic fibroblasts exhibited increased Plk4 hypermethylation and downregulation along with increased centrosome numbers and multinucleation.
These results suggest that aberrant Plk methylation is correlated with the development of HCC in mice.
The Polo-like kinases (Plks) are a highly conserved family of serine-threonine kinases, found from unicellular eukaryotic organisms to higher multicellular eukaryotes. The mammalian Plks (Plk1-4) have been shown to play major roles in cell cycle regulation, centrosome dynamics and the cellular response to stress. Furthermore, perturbations in individual Plk protein levels have been associated with malignancies. For example, high levels of Plk1 are indicative of a poor prognosis in esophageal, non-small cell lung cancer and oropharyngeal carcinomas [1, 2] and have been observed in various forms of cancers including gastric, breast, ovarian, endometrial, gliomas, thyroid and melanomas . In contrast, Plk3 is downregulated significantly in carcinomas of the lung, head and neck [4, 5]. The Plk2 gene is downregulated in lymphomas and B-cell malignancies . In the case of Plk4, over 50% of aged Plk4 heterozygous (Plk4 +/-) mice develop tumours in comparison to only 3% of their wild-type littermates, the major site of tumour formation being the liver and lung . In mice, Plk4 is haploinsufficient for tumour suppression, while in humans, loss of heterozygosity (LOH) for the Plk4 gene was found in 60% of a small sample of human hepatocellular carcinomas (HCC) cases. The increased rate of tumourigenesis is likely related to the generation of aneuploidy, as altered Plk4 levels result in abnormal centrosome numbers , furthermore Plk4 may also play a key role in a DNA damage response pathway consistent with its phosphorylation of p53 , and Chk2 . In general, overexpression of Plk1 is typically considered to be oncogenic in nature while the remaining Plks likely function as tumour suppressors.
Recently it has become evident that the hypermethylation of CpG islands of tumour-suppressor genes, histone modification and chromatin remodelling are common events in cancers (for review see ). Individual Plk gene epigenetic modifications associated with malignancy have previously been documented for Plk2 where its methylation-dependent silencing was detected at a high rate in B-cell malignancies and Burkitt's Lymphoma as well as in follicular lymphoma [11, 12]. The correlation between the methylation status of the Plks and malignancy has not been studied in detail. In this regard, as noted below, we initially identified a gender disparity for the development of HCC in Plk4 +/- mice. Previously, the development of HCC was attributed to haploinsufficiency for Plk4 rather than via loss of heterozygosity . Given that there is accumulating evidence that epigenetic changes are a driving force in the development of HCC , we were interested in determining whether a relationship exists between individual Plk epigenetic modifications in the context of Plk4 haploinsufficiency and the development of HCC.
Lower Plk4 levels likely play a role in malignancy by affecting genomic stability through a mechanism related to Plk4's role in centrosome duplication  and/or DNA damage pathways . We therefore examined the levels of Plk4 transcripts and found that the levels were substantially lower in males versus female mice as early as 9 months of age (Figure 1e) and were greater than 10 fold lower in livers and liver tumours from aged Plk4 +/- mice compared to wild type males and females and Plk4 +/- females (Figure 1f). Similarly, Plk4 protein was also significantly reduced in tumours (Figure 1g). It is noted that, while livers from Plk4 +/- mice were grossly normal, they displayed variable amounts of Plk4 transcripts with an average that is significantly lower than that found in Plk4 +/+ mouse livers. Similarly, at the protein level, in Plk4 +/-, we see varied amounts. It is noted that the Plk4 +/- mice typically develop HCC 18-24 months on with some cases as early as 13 months. We propose that this likely reflects varying stages of progression towards the development of HCC; suggesting that reduced levels of Plk4 as a result of promoter methylation may precede the appearance of visible tumours. Low levels of Plk4 have been shown to result in the generation of mono-polar spindles and aneuploidy in both cell lines and tissues [7, 8]. This exemplifies the possibility that epigenetic modifications may play a role in gender biases for malignancy and corresponds to our observation that epigenetic modifications of the Plk4 gene leads to further Plk4 downregulation, particularly in males.
In order to determine if Plk4 methylation status is correlated with the development of HCC in humans, we also examined a limited number of human liver samples (See Additional File 2). We found that in normal human hepatic tissue the Plk4 promoter region was not methylated in samples taken from patients with no history of HCC. In the case of HCC samples, we detected Plk4 CpG island hypermethylation and downregulation of Plk4 transcript levels as well as barely detectable methylation of the Plk1 promoter region. In 3 of 6 samples we found that the corresponding Plk1 transcript levels were higher than in the normal control (Additional File 2e). We did not detect any changes for Plk2 and Plk3 promoter methylation (data not shown). Since we began this aspect of our study, Pellegrino et al. (2010) examined a large cohort of human HCC samples and reported Plk2-3 downregulation in human hepatocellular carcinoma correlated with either promoter hypermethylation and/or loss of heterozygosity at the Plk2-3 loci . In the case of Plk4, many of the samples displayed loss of heterozygosity with no methylation within the Plk4 promoter region. They did not report any analysis for the methylation status of Plk1. Their inability to detect methylation changes for Plk4 and ours for Plks2-3 may be a reflection of the use of different primers for methylation specific PCR (MSP) (Additional file 3) which samples a small subset of the potentially methylated residues within a CpG island. Together these results suggest that in general, epigenetic changes within the Plks may contribute to malignancy in humans.
In general, global hypermethylation increases with age; however, studies on aberrant methylation of genes associated with HCC, like in many other malignancies, are characterized by an overall general increase in global hypomethylation along with increased rates of hypermethylation of tumour suppressors . We employed an ELISA-based assay (Epigentek) in order to quantitatively measure genomic methylation. Interestingly, we found no significant difference between the 9 month old wild type males and age-matched wild type and Plk4 +/- females (Figure 3e). However, consistent with what has been shown with age progression, we found an overall increase in the global methylation of genomic DNA in wild type male mice and both Plk4 wild type and heterozygous female mice from 9 to 20 months. In contrast, there was a decrease in global methylation in Plk4 +/- male mice over the same time period (*p < 0.05). Furthermore, significantly higher levels of global methylation were found in young Plk4 +/- male mice compared to their wild type littermates (**p < 0.001), while the opposite is true for the Plk4 +/- female mice, where they had significantly lower levels of global methylation compared to young wild type females (***p < 0.05). Although, as the females age, both genotypes have similar levels of global methylation. These results suggest that there is an interplay between gender and Plk4 haploinsufficiency that affects global methylation in liver tissue.
p53 has also been found to be an upstream negative regulator of Plk4 via histone deactylation (HDAC) . We therefore examined p53 levels in normal and tumour tissue and found that both p53 and p21 were upregulated in tumour tissue compared to the normal tissue (Figure 3f). p53 is also a substrate for Plk4  and p53 levels/activity are upregulated as a result of haploinsufficiency in MEFs . These observations suggest that increased p53 levels/activity, a consequence of Plk4 haploinsufficiency, may also contribute to repressive chromosome structure and the reduced transcript profiles seen in aged and tumourigenic Plk4 +/- mice.
Alcohol has become an emerging environmental player in the modification of the epigenome . In humans, chronic alcoholism has been shown to increase availability of blood homocysteines, which in turn modify s-adenosyl methyltransferase (MATs) levels, an enzyme responsible for the transfer of methyl groups to DNA. Furthermore, these patients showed a significant increase in global DNA methylation by up to 10% . There is increasing evidence that alcohol consumption, a known risk for the development of HCC, can increase the methylation status of promoters with a subsequent decrease in gene expression [24–26]. In liver cells, the presence of alcohol results in an increase in the formation of reactive oxygen species, which are in turn responsible for hepatocyte damage, cellular apoptosis, and the tumour promoting effect of ethanol . Interestingly, we have preliminary evidence of increased Plk4 methylation in human cirrhotic livers with no evidence of viral infection (see Additional File 2). This, coupled with the associated correlation between alcoholism and HCC development led us to examine the methylation status and expression of the individual Plks in a cell-based model of chronic ethanol exposure.
Unlike mutations or deletions that lead to the aberrant expression of tumour suppressor genes, epigenetic modifications, like DNA methylation, are reversible via the use of hypomethylating drugs that inhibit DNA methyltransferase activity and/or inhibit HDACs . Concurrent alcohol and epigenetic drug treatments revealed that 5-aza-2'-deoxycytidine, a DNA hypomethylator, and valproic acid, which has been shown to be an HDAC inhibitor, partially restored Plk4 transcript levels, while no significant differences were seen with trichostatin A (an HDAC inhibitor) treatment (Figure 4e).
Modification of the methylation status and corresponding expression levels of both Plk4 and Plk1 are likely contributing factors in the development of HCC in both mice and humans. This creates interesting possibilities in that epigenetic modifications are potentially reversible through the use of demethylating and HDAC inhibiting drugs as both prophylactic and therapeutic tools. This may lead to the development of novel treatment options for HCC.
We determined that a gender disparity exists for the development of HCC in the Plk4 mouse model. This disparity was correlated with increased DNA methylation at the Plk4 locus and higher risk of developing hepatocelluar carcinomas in aged male Plk4 heterozygous mice as compared to female mice. In contrast, we discovered the opposite correlation for Plk1 where in normal liver tissue the Plk1 promoter is hypermethylated while in tumours, Plk1 CpG islands become hypomethylated and the gene upregulated. This represents a novel form of regulation for Plk1 that may have implications for its expression in other tumour types. Furthermore, we determined that chronic alcohol exposure, well known to be implicated in the development of cirrhosis leading to HCC, also leads to Plk4 promoter hypermethylation and downregulation, accompanied by defects in the control of centrosome numbers and by the occurrence of multinucleation in cells. Aberrant Plk4 methylation and expression in chronically exposed MEFs could be rescued by treatment with known hypomethylating and/or HDAC inhibiting drugs.
Mouse primer sequences
5'aca aac acc tct ttt ata tct aca tc 3'
5'tgg ttt gag tat tag ttg att ttg g 3'
5'acg aac acc tct ttt ata tct acg tc 3'
5'gtt ggt tcg agt att agt cga ttt c 3'
5' caa act tta ccc aaa acc tac tcac 3'
5'ata ggg tta gtt tgg atg ttt gtt t 3'
5' aaa ctt tac cca aaa cct act cg 3'
5'ggt tag ttc gga cgt ttg ttc 3'
5'cac act ctc cac ttc tta aaa aca a 3'
5' att tta tta tta gtg ttt gtg tta tgg 3'
5'aca ctc tcc act tct taa aaa cga a 3'
5' aat tta tta tta gcg ttc gcg tta c 3'
B1 Element U
5'-taa cct caa act caa aaa tcc acc-3'
5'gtt ggg tgt agt ggt ata tat ttt taa ttt ta 3'
B1 Element M
5' gtc ggg cgt agt ggt ata tat ttt t 3'
All murine samples were obtained from our breeding colony, with all protocols for animals approved by the University of Windsor Animal Care Committee according to the Canadian Council on Animal Care guidelines. Plk4 +/- mutant mice on a 129Sv/CD1 background were obtained as described  and backcrossed with CD1 mice to establish a colony of Plk4 wild type and Plk4 heterozygous littermates. Mice were maintained under normal light cycle and on regular chow. All tissue samples were obtained from aged matched littermates. For murine hepatocellular carcinoma (HCC) samples, it is noted that Plk4 +/- mice develop a high rate of liver and lung tumours by 18-24 months of age  and thus the analysis was performed on spontaneously occurring hepatocellular carcinomas.
Mouse embryonic fibroblasts (MEFs) were harvested from Plk4 wild type CD1 mice at day 12.5 post coitum as described previously in  and cultured with DMEM supplemented with 20% FBS (Sigma), 1% penicillin G sodium 10,000 U/mL, streptomycin sulphate 10,000 ug/mL, and gentamycin 10 mg/mL.
Protein from fresh tissue was extracted using the Trizol reagent (Invitrogen) according to manufacturer's provided protocols. Cell lysates were obtained from cells treated with buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.5% Triton X with EDTA free protease inhibitor cocktail tablets (Roche). Western blot analysis was performed using 20 ug of total protein. Primary antibodies were as follows: p53 (Sigma), Plk1 (Abcam), p21, Plk2, Plk3 (Santa Cruz), Plk4, GAPDH (Cell Signaling), and Actin (Sigma). Secondary antibodies were anti-rabbit (Amersham) and anti-mouse horseradish peroxidase (HRP) (Sigma).
RNA was extracted from cells and tissues using the RNAeasy kit (Qiagen) according to manufacturer's recommendations. cDNA was generated using the "First Strand cDNA synthesis kit" according to the manufacturer's instructions. Quantitative real time PCRs (qPCR) were conducted in an ABI 7300 instrument using 250 ng of cDNA with TaqMan Gene Expression Assays (Applied Biosystems) for mouse Plk1 and Plk4. Rodent GAPDH probe was used as an internal control. Relative quantity (RQ) values were generated by the ABI 7300 system SDS software. The error bars represent the upper and lower limit of the standard error from the mean expression level (RQ) as analyzed by the SDS software. The error bars are calculated based on 95% confidence limits.
MEF cells were fixed in 3.7% paraformaldehyde and probed with a mouse γ-tubulin primary antibody (Sigma) followed by an anti-mouse alexa fluor 568 secondary antibody (Invitrogen). The cells were then briefly incubated in Hoescht 33342. Cells were analyzed with a Zeiss Axioskop 2 mot plus microscope and Northern Eclipse imaging software. Conditions for immunofluorescence were as described previously .
Wild type MEFs were exposed to 25 mM or 50 mM ethanol per day for 7 days. Trichostatin A, 5 aza-2'-deoxycytidine, and valproic acid were administered concurrently at concentrations of 1 nM, 10 nM, and 0.5 mM respectively.
We thank A. Kozarova, A. Swan, and S. Varmuza for helpful discussion. This work was supported by a grant from the National Science and Engineering Research Council of Canada (NSERC) (J.W.H.) and the University of Windsor.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.