Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS One. 2012; 7(3): e34329.
Published online 2012 Mar 26. doi:  10.1371/journal.pone.0034329
PMCID: PMC3312909

FOXM1 Induces a Global Methylation Signature That Mimics the Cancer Epigenome in Head and Neck Squamous Cell Carcinoma

Andrew Yeudall, Editor


The oncogene FOXM1 has been implicated in all major types of human cancer. We recently showed that aberrant FOXM1 expression causes stem cell compartment expansion resulting in the initiation of hyperplasia. We have previously shown that FOXM1 regulates HELLS, a SNF2/helicase involved in DNA methylation, implicating FOXM1 in epigenetic regulation. Here, we have demonstrated using primary normal human oral keratinocytes (NOK) that upregulation of FOXM1 suppressed the tumour suppressor gene p16INK4A (CDKN2A) through promoter hypermethylation. Knockdown of HELLS using siRNA re-activated the mRNA expression of p16INK4A and concomitant downregulation of two DNA methyltransferases DNMT1 and DNMT3B. The dose-dependent upregulation of endogenous FOXM1 (isoform B) expression during tumour progression across a panel of normal primary NOK strains (n = 8), dysplasias (n = 5) and head and neck squamous cell carcinoma (HNSCC) cell lines (n = 11) correlated positively with endogenous expressions of HELLS, BMI1, DNMT1 and DNMT3B and negatively with p16INK4A and involucrin. Bisulfite modification and methylation-specific promoter analysis using absolute quantitative PCR (MS-qPCR) showed that upregulation of FOXM1 significantly induced p16INK4A promoter hypermethylation (10-fold, P<0.05) in primary NOK cells. Using a non-bias genome-wide promoter methylation microarray profiling method, we revealed that aberrant FOXM1 expression in primary NOK induced a global hypomethylation pattern similar to that found in an HNSCC (SCC15) cell line. Following validation experiments using absolute qPCR, we have identified a set of differentially methylated genes, found to be inversely correlated with in vivo mRNA expression levels of clinical HNSCC tumour biopsy samples. This study provided the first evidence, using primary normal human cells and tumour tissues, that aberrant upregulation of FOXM1 orchestrated a DNA methylation signature that mimics the cancer methylome landscape, from which we have identified a unique FOXM1-induced epigenetic signature which may have clinical translational potentials as biomarkers for early cancer screening, diagnostic and/or therapeutic interventions.


Understanding the epigenetic mechanism regulating stem-cell fate determination provides fundamental insights into the physiology of tissue regeneration and pathogenesis of cancers. The best studied epigenetic mechanism perturbed during cancer initiation and progression is DNA methylation which chemically adds methyl groups to cytosines at their 5′ positions, predominantly at CpG dinucleotides in the mammalian genomic DNA [1]. DNA methylation involves three key DNA methyltransferases: DNMT1, DNMT3A and DNMT3B. DNMT1 has classically been implicated in maintenance of existing methylated DNA, whereas, DNMT3A and DNTM3B in de novo DNA methylation [1]. The heritable nature of DNA methylation enables cells to determine cell potency/fate without changing the primary sequence of genomic DNA. The reversibility of DNA methylation programming renders cell fate specification highly plastic and reversible. Epigenetic reprogramming involving changes in DNA methylation has been implicated in all stages of cancer evolution [2], [3]. It has also been shown that epigenetic reprogramming precedes the initiation of cancer-like stem/progenitor cells [4]. It is now well-accepted that cancer cells exploit the reversible and heritable properties of DNA methylation to perturb the balance between stem/progenitor cell renewal and differentiation thereby promoting cancer initiation and progression [2], [3], [4].

FOXM1 (isoform B) was first found to be a downstream target of an oncogenic Sonic Hedgehog signalling pathway via a glioma family zinc finger transcription factor 1 (Gli1) in basal cell carcinomas [5]. Subsequent studies revealed that FOXM1 was ubiquitously upregulated in the majority of human cancers [6], [7] which include brain, liver, breast, lung, stomach, pancreas, colon, kidney, bladder, prostate, testis, ovary, uterus, cervix, blood (acute myeloid leukaemia), cutaneous melanoma, head and neck squamous cell carcinomas [8], [9].

In the quest to understand the oncogenic mechanism of FOXM1, we have recently shown that FOXM1 induces cancer initiation by promoting adult human epithelial stem/progenitor cell renewal and by antagonising differentiation [10]. Others have demonstrated that FOXM1 plays a key role in maintaining stem/progenitor cell renewal through pluripotency genes including Oct4, Nanog, Sox2 and Bmi1 [11], [12]. Our previous work identified a FOXM1 downstream target HELLS [8], a human embryonic stem cell factor/lymphoid-specific SNF2/helicase involved in chromatin remodelling and DNA methylation [13], [14], implicating FOXM1 in epigenetic regulation during stem/progenitor cell renewal [8], [10]. However, it was unclear whether FOXM1 has a role in epigenetic regulation. In this study, using primary normal human oral keratinocytes and head and neck squamous cell carcinoma (HNSCC) tumour cell lines and tumour biopsy tissues, we investigated the role of FOXM1 in the regulation of gene promoter methylation at both single gene and genome-wide levels. This led to the first evidence in normal primary human oral epithelial cells that FOXM1 induces a methylation landscape resembling a cancer epigenome found in HNSCC tumour tissues.


Clinical Tissues

The use of human tissue in this study has been approved by our host institutions (Barts & the London NHS Trust and the School of Medicine & Dentistry, Queen Mary University of London) and the UK National Research Ethics Committee. All clinical samples, which were surplus to diagnosis, were collected according to local ethical committee-approved protocols and written informed patient consent was obtained from all participants. Pairs of normal margin and HNSCC tumour core tissue biopsies were histopathological pre-validated by our collaborating pathologists prior to use for this study. Fresh biopsy tissue samples were preserved in RNALater (Cat# AM7022, Ambion, Applied Biosystems, Warrington, UK) and stored short-term at either 4°C (1–2 days) or −20°C (up to 1 week) prior to transportation and subsequent long-term storage at −80°C until use.

Cell culture

All primary normal human oral keratinocytes (OK355, HOKG, OK113, NOK, NOK1, NOK3, NOK16 and NOK376) were extracted from normal oral mucosa tissues donated by healthy disease-free individuals undergoing wisdom tooth extraction and cultured as previously described [8], [15]. Oral dysplastic precancer cell lines (OKF6/T [16], POE9n [17], DOK [18], D19 [19], D20 [19]) and oral SCC cell lines (SCC4 [20], SCC9 [20], SCC15 [20], SCC25 [20], SqCC/Y1 [21], UK1 [22], VB6 [22], CaLH2 [22], CaDec12 [22], 5PT [22], H357 [22]), SVpgC2a [23] and SVFN1-8 [8] were all well-established cell lines cultured as described previously [8], [10], [15].


Protein extraction and separation on SDS-PAGE gels and immunoblotting was performed as previously described (5). A mouse monoclonal antibody for p16INK4A (1∶2000 dilution; Cat# 551154, BD Biosciences) and a rabbit polyclonal anti-GAPDH (1∶20,000 dilution; Cat# 9485, Abcam) were used for immunoblotting.

RNA interference

Pre-validated gene-specific siHELLS (ON-TARGETplus SMARTpool HELLS, Cat# L-017444-09,10,11,12), control siCTRL (ON-TARGETplus Non-targeting Pool, Cat# D-001810-10-05) and siRNA transfection reagent (DharmaFECT 1, Cat# T-2001-02) were purchased from Dharmacon, Thermo Fisher Scientific. An initial dose-response experiment was performed according to manufacturer's instructions to determine the optimum transfection efficiency. siRNA at 10 nM (48-hour incubation) was found to be the optimum final concentration which was therefore used in all subsequent experiments. The effect of gene silencing was validated by quantification of the target gene mRNA expression (HELLS) by absolute reverse transcription qPCR.

Retroviral transduction

Retroviral supernatant and transduction procedures were performed using our established protocols [8], [10], [15]. Equal levels of EGFP and FOXM1 (isoform B) expression were achieved by serial retroviral supernatant titration experiment and subsequently EGFP plasmid copy number confirmed by qPCR using genomic DNA extracted from transduced cells according to our previously established method [15]. The levels of ectopic FOXM1 expression in the primary keratinocytes were titrated to replicate levels found in cancer cells as reported previously [8], [10], [15] (see Figure 1C). Transduced cells were cultured for 3–5 days to allow transgene expression prior to experiment.

Figure 1
Upregulation of FOXM1 suppressed p16INK4A expression in primary human oral keratinocytes.

Nucleic Acids Preparations from Tissues and Cells

All tissue biopsies were digested by proteinase K (Cat# 03115887001, Roche Diagnostics Ltd., England, UK) prior to simultaneous mRNA extraction (Dynabeads mRNA Direct kit, Cat# 610.12, Invitrogen) and genomic DNA (gDNA) extraction (by standard phenol∶chloroform method on mRNA-depleted lystates). mRNA was immediately reverse transcribed into cDNA (Transcriptor cDNA Synthesis kit, Cat# 04897030001, Roche Diagnostics). gDNA were fragmented by MseI digestion (37°C, 16 h) prior to enrichment for CpG-methylated DNA using a MBD2b/MBD3L1-conjugated magnetic bead-based system according to manufacturer's protocol (MethylCollector Ultra kit, Cat# 55005, Active Motif Europe, Belgium).

Genome-wide Promoter Methylation Profiling

According to manufacturer's protocol and requirements, input MseI-digested gDNA and methylation-enriched DNA from each cell sample (NOKG, NOKF and SCC15) were amplified to generate 6 µg DNA using WGA2 GenomePlex (Sigma) prior to microarray experiments performed by Roche NimbleGen microarray service using Human DNA Methylation 3x720K CpG Island Plus RefSeq Promoter Array (Cat# 05 924 600 001; NimbleGen System, Reykjavik, Iceland) based on genome built HG18, with promoter upstream/downstream tilling of −2.44/+0.61 kb, covering a total of 27,728 CpG islands across the whole genome (GEO Platform: GPL14361). Microarray data generated in this study is MIAME compliant and has been deposited in a MIAME compliant database at Gene Expression Omnibus repository (GEO Series accession number: GSE31767).

Real-time absolute quantitative PCR

Standard curve-based real-time absolute quantitative PCR were performed using SYBR Green I Master (Cat# 04887352001, Roche Diagnostics Ltd, England, UK) in the 384-well LightCycler 480 qPCR system (Roche) according to our established protocols [8], [9], [15] which are MIQE compliant [24]. Methylation-specific PCR conditions were performed as described previously [25], [26]. All primers used in this study are listed in Figure S1. Previously validated isoform B-specific FOXM1 primers were used to specifically quantify FOXM1 (isoform B) mRNA expression in this study [8]. All target genes were normalised to two stable reference genes (YAP1 and POLR2A) previously validated to be amongst the most stable reference genes across a wide variety of primary human oral cells, dysplastic and HNSCC cell lines [8].

Results and Discussion

Given our previous finding that FOXM1 (isoform B) promoted stem/progenitor cell renewal through perturbing the differentiation pathway [10], we initially questioned the involvement of a tumour suppressor gene p16INK4A (CDKN2A) given that it has been shown to regulate epithelial stem/progenitor cell differentiation [27] and it is the most commonly inactivated gene in cancer [28]. Here, we showed that ectopic FOXM1 expression suppressed both mRNA and protein expression of p16INK4A in primary human oral keratinocytes (Figure 1A). Unfortunately, as reported previously silencing endogenous FOXM1 expression causes cell cycle arrest [29] which precluded further experiments using RNAi on the notoriously sensitive primary human oral keratinocytes [8], [10]. Nevertheless, our FOXM1 overexpression experiments conclusively showed that FOXM1 upregulation suppressed p16INK4A gene expression in primary human oral keratinocytes. This is in agreement with previous findings that FOXM1 suppresses the senescence pathway mediated by p16INK4A in cancer cells [30].

Inactivation of p16INK4A gene expression could be a result of a number of mechanisms including gene deletion and promoter hypermethylation. Given that FOXM1 targets HELLS which regulates DNA methylation [13], [14], we hypothesised that FOXM1 may be suppressing p16INK4A expression through promoter hypermethylation via HELLS. To test this, we knockeddown HELLS by siRNA in an HNSCC cell line SVFN5, a FOXM1-induced transformed oral buccal keratinocyte SVpgC2a line [8], that expresses high levels of endogenous HELLS and low levels of p16INK4A. This causes re-activation of the mRNA expression of p16INK4A (Figure 1B) and concomitant downregulation of two DNA methyltransferases DNMT1 and DNMT3B but no effect on DNMT3A expression. The fact that p16INK4A inhibition could be reactivated argues against gene deletion as a mechanism for p16INK4A inactivation. Our results are consistent with previous findings that HELLS interacts with DNMT1 and DNMT3B [31] to suppress p16INK4A gene expression [32] through epigenetic modifications.

To further validate that the expression of FOXM1, HELLS and p16INK4A genes correlate with cancer progression and whether there are any associations with genes involved in DNA methylation, we measured the endogenous mRNA expression levels of FOXM1, p16INK4A, HELLS, BMI1, involucrin (IVL, a differentiation marker has been shown to be negatively regulated by FOXM1 [10]) and 3 key DNA methyltransferases (DNMT1, DNMT3A, DNMT3B) in a panel of 24 cell strains/lines consisting of 8 strains of primary normal human oral keratinocytes (from normal oral mucosa tissues), 5 dysplasia and 11 HNSCC cell lines.

In agreement with previous findings [8], [10], FOXM1 showed dose-dependent upregulation during tumour progression from dysplasia to HNSCC (Figure 1C). Across the panel of 24 cell strains/lines, we have found that the endogenous mRNA expression of FOXM1 correlated inversely with p16INK4A but correlation efficiency was weak (R2 = 0.23, Figure 1D). The downregulation of p16INK4A expression was found to be more pronounced in dysplastic compared to HNSCC cell lines. Such p16INK4A expression pattern is in complete agreement with in vivo p16INK4A protein expression pattern found in oral dysplasia and SCC tissues [33]. Consistently, BMI1, a polycomb group oncogene which is an upstream regulator of p16INK4A gene [34] and also a downstream target of FOXM1 [12], [30], showed positive co-expression with FOXM1 (R2 = 0.64, Figure 1E) but weak inverse correlation with p16INK4A (R2 = 0.42, data not shown) supports the evidence that p16INK4A expression is independently regulated by BMI1 during oral carcinogenesis [35]. The discordant expression levels between FOXM1 and p16INK4A in cancer cells may be due to the fact that p16INK4A can be deregulated through a number of different mechanisms, such as inactivating mutation (may result in upregulation due to feedback mechanism), gene deletion, gene amplification (of functional gene but defective downstream signalling), promoter hypermethylation, etc. This may result in varying p16INK4A expression independent of FOXM1 levels in the “cancer” cell lines. Hence, whilst FOXM1 can induce promoter hypermethylation of p16INK4A in “normal” cells, such effect may be perturbed in “cancer” cells.

Expression of DNMT1 (R2 = 0.84; Figure 1H) and DNMT3B (R2 = 0.89; Figure 1J), but not DNMT3A (R2 = 0.13; Figure 1I), showed significant positive co-expression with FOXM1 which are in agreement with our findings above (Figure 1B) that silencing the FOXM1-downstream target HELLS led to concomitant downregulation of DNMT1 and DNMT3B but no effect on DNMT3A expression. It is unclear why DNMT3A was not affected. Published literature indicates that although both DNMT3A and DNMT3B are involved in de novo methyltransferase activity, they serve non-overlapping roles [1]. Nevertheless, the involvement of both DNMT1 and DNMT3B implicates a role for FOXM1 and HELLS in triggering both maintenance and de novo DNA methylation activities [1]. Expectedly, HELLS were positively (R2 = 0.76, Figure 1F) and IVL were negatively (R2 = 0.52, Figure 1G) correlated with FOXM1 as shown previously [8], [9], [10]. Collectively, these results provide the first evidence in human cells that FOXM1 may be acting through HELLS, DNMT1 and DNMT3B to suppress p16INK4A gene expression. Given that HELLS, DNMT1 and DNMT3B have been previously shown to modulate p16INK4A promoter methylation [31], [32], we hypothesised that FOXM1 may be triggering p16INK4A gene silencing through promoter hypermethylation.

To investigate promoter CpG DNA methylation, we quantified the level of p16INK4A promoter methylation using bisulfite modification and methylation-specific quantitative PCR (MS-qPCR; Figure 2A and Figure S1). Overexpression of FOXM1, but not EGFP, was found to induce p16INK4A promoter hypermethylation (P<0.05) which was significantly reversed (P<0.001) by a DNA demethylating agent 5-aza-2′-deoxycytidine (5Aza) in primary human oral keratinocytes (Figure 2B). These results confirmed a role of FOXM1 in suppressing p16INK4A expression through promoter hypermethylation. In support for FOXM1 in initiating oncogenesis through the inhibition of p16INK4A, it has been shown that epigenetic silencing of p16INK4A induces cellular immortalisation in mouse embryonic fibroblasts [36]. Furthermore, our previous finding that FOXM1 expression co-expressed with an epithelial stem cell marker ΔNp63α in the proliferating stem/progenitor oral keratinocyte subpopulation [10], and that ΔNp63α has been shown to target HELLS to induce squamous cell carcinoma formation in mice [37], together suggest a possible role for FOXM1 (via HELLS) in triggering oncogenesis through silencing p16INK4A. The exact oncogenic mechanism is beyond the scope of this study. Nevertheless, our current data providing the first evidence that FOXM1 is able to induce promoter hypermethylation at a single gene level offers a glimpse of possibility that aberrant upregulation of FOXM1 may perturb the epigenetic regulation of DNA methylation at genome-wide level.

Figure 2
FOXM1 induces promoter hypermethylation of p16INK4A gene in primary human oral keratinocytes.

We and others have previously established a central role for FOXM1 in the maintenance of genome stability whereby aberrant FOXM1 expression causes global genomic instability [8], [15], [38]. Furthermore, the findings that FOXM1 targets an epigenetic/stem cell modulator HELLS during cancer initiation [8], [14] and FOXM1 directly induces p16INK4A promoter hypermethylation (Figure 2) prompted us to hypothesise that aberrant upregulation of FOXM1 perturbs the methylome. To test this hypothesis, we performed a non-bias genome-wide promoter methylation microarray profiling on primary normal oral human keratinocytes (NOK) either overexpressing a control gene EGFP (NOKG) or FOXM1 (NOKF) (see Figure 1C for FOXM1 gene expression levels of NOKG and NOKF cells), and also on an HNSCC cell line (SCC15). SCC15 was chosen in this study as a positive control because the promoter of p16INK4A gene (CDKN2A) has been previously shown to be hypermethylated and could be reactivated by 5Aza [39], hence allowing us to validate the methylation array data. FOXM1 was found to induce a global hypomethylation pattern similar to that found in the HNSCC cell line, compared to control NOK cells expressing EGFP (Figure 3A). Comparing the methylation patterns by regression correlation analyses amongst the three cell types (NOKG, NOKF and SCC15), only NOKF vs SCC15 gave a positive correlation pattern, whereas NOKF or SCC15 each produced an inverse correlation with the control NOKG (Figure 3B). This indicates that overexpression of FOXM1, but not EGFP, induces a methylation landscape similar to that found in SCC15. Both global hypomethylation and focal hypermethylation (affecting individual genes) are typical methylation patterns found in cancer [2], [3]. The fact that upregulation of FOXM1 induces these methylation patterns in “normal” cells indicates that aberrant expression of FOXM1 is changing the methylation landscape towards those of cancer. The consequence of global hypomethylation has been shown to cause genomic instability [2], [3], this may provide a mechanism for our previous findings that aberrant FOXM1 expression causes genomic instability in primary normal human keratinocytes [8], [15]. Although global hypomethylation appeared to be the dominating effect, it has been shown that focal hypermethylation silencing key tumour suppressor genes (eg. p16INK4A) also plays important role in oncogenesis [2], [3].

Figure 3
Upregulation of FOXM1 (isoform B) induces a global shift in methylation pattern that mimics the cancer epigenome.

To validate our hypothesis that FOXM1-orchestrated a methylation signature that mimics a cancer methylome, differentially methylated genes (100 most hypomethylated and 100 most hypermethylated) were initially selected for inverse comparisons between NOKG and NOKF/SCC15, and a subset of 30 consensus genes, shared between NOKF and SCC15 cells, with opposing methylation status to NOKG control cells, were subsequently shortlisted for further analyses (Figure 3C). If these candidate FOXM1-induced differentially methylated genes were indeed an epigenetic signature of cancer, we hypothesised that HNSCC tumour tissues should retain an inverse in vivo mRNA expression signature of these candidate genes. To verify this, we performed absolute qPCR to quantify each of the 30 candidate genes: i, the relative levels of promoter DNA methylation of each gene in NOKG vs NOKF cells, and, ii, the relative mRNA expression levels in paired normal margin vs HNSCC tumour tissue samples. Correlation regression analyses of the 30 candidate genes showed an inverse relationship (R2 = 0.62; Figure 3D, left panel) between gene expression of HNSCC tumour tissues and DNA methylation of NOKF cells.

Interestingly, hypomethylated genes showed significantly higher inverse correlation pattern (R2 = 0.92; Figure 3D, right panel) than the hypermethylated genes (R2 = 0.27; Figure 3D, middle panel). This suggests that promoter hypomethylation exhibited a stronger effect on transcriptional activation compared to promoter hypermethylation on transcriptional repression. One explanation could be that it may be easier to detect transcriptional activation following promoter hypomethylation as opposed to detecting transcriptional repression which depends on whether the genes were activated prior to hypermethylation. Our results indicate that hypo/hypermethylation may not be a simple symmetrical on/off switch for gene transcription. Further studies are required to delineate the transcriptional mechanisms regulated by promoter DNA methylation/demethylation.

Of the list of 15 novel FOXM1-induced hypermethylated genes (Figure 3D, middle panel), 4 genes (C6orf136, MGAT1, NDUFA10 and PAFAH1B3) had significantly downregulated mRNA expression levels in HNSCC tumours, along with the positive control p16INK4A (CDKN2A). Little published gene information was available for C6orf136. MGAT1 [mannosyl (alpha-1,3-)-glycoprotein beta-1,2-N-acetyl-glucosaminyltransferase] has been implicated in glycerolipid metabolism [40]; NDUFA10 (NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 10, 42 kDa) in mitochondrial metabolism [41] and PAFAH1B3 (platelet-activating factor acetylhydrolase 1b, catalytic subunit 3, 29 kDa) in brain development [42] and spermatogenesis [43]. Given that their gene expressions were suppressed through promoter hypermethylation in tumour tissues, we speculate that they may be tumour suppressor genes. However, their roles in tumourigenesis remained to be investigated.

Of the list of 14 novel FOXM1-induced hypomethylated genes (Figure 3D, right panel), 4 genes (SPCS1, FLNA, CHPF and GLT8D1) had significantly upregulated mRNA expression levels in HNSCC tumours. FLNA (filamin A, alpha), an actin-binding protein involves in cytoskeletal/membrane remodelling and cellular motility [44], [45], has been implicated in melanomagenesis [44], [46], prostate [47], [48], [49], breast [50], lung [51], liver [52] and ovarian cancers [53]. CHPF (chondroitin polymerizing factor), involved in extracellular matrix regulation [54], has recently been implicated in colorectal cancer [55]. SPCS1 (signal peptidase complex subunit 1 homolog) and GLT8D1 (glycosyltransferase 8 domain containing 1) are located adjacent to each other at chromosome 3p21.1. Given that their gene expressions were upregulated in tumour tissues, we speculate that they may be oncogenes. However, their roles in oncogenesis remained to be investigated.

Collectively, these results confirmed that aberrant expression of FOXM1 triggers genome-wide methylomic alterations that mimic the in vivo cancer methylome of HNSCC tumour tissues. We speculate that this may be a mechanism exploited by FOXM1 to induce progenitor/stem cells expansion [10] through methylome reprogramming to antagonise differentiation.

In summary, we have shown for the first time that aberrant upregulation of a single oncogene FOXM1 in primary normal human oral epithelial cells orchestrated a cancer-like methylome landscape, from which we have identified a unique set of FOXM1-induced differentially methylated genes. We further provided evidence that their in vivo gene expression signatures were retained in HNSCC tumour tissues. Given that epigenetic alteration precedes gene expression, we speculate that the FOXM1-induced differentially methylated genes have strong potential as epigenetic biomarkers for early cancer screening, diagnostic, prognostic and/or therapeutic interventions.

Supporting Information

Figure S1

Absolute qPCR primers. (A) Nucleotide sequence of the bisulfite treated promoter region of p16INK4A and their respective primer sequences used in this study. Details of qPCR conditions were performed according to published methods [25], [26]. (B) qPCR primer sequences of the 30 candidate FOXM1-induced differentially methylated genes. Colour shaded loci indicate that the genes were adjacent or nearby. Promoter CpG islands (CGI) for each gene are annotated as either ‘S’ (sense strand), ‘AS’ (antisense strand) or ‘-’ (no CGI within promoter region). All primer pairs produce a single melting peak. Standard curves were generated for each gene for absolute quantification of unknown samples according to protocols described previously [8].



We thank Professors Ian Mackenzie (Centre for Cutaneous Research) and Ken Parkinson (CDOS) for their critical advice and precious gifts of various HNSCC cell lines. We thank Dr Daniela Costea (The Gade Institute, University of Bergen, Norway) for her precious primary human oral keratinocyte strains. We also thank Mr Muy-Lip Teh (Air Operations Software Engineering Department, Thales Australia Centre, Melbourne, Australia) for his advice on data processing and use of the R Environment for statistical analysis.


Competing Interests: The authors have declared that no competing interests exist.

Funding: This work was co-funded by the Wellcome Trust (MTT, EG), the Facial Surgery Research Foundation - Saving Faces (MTT, AW) and the Institute of Dentistry (DP, RT, AN), Barts and the London School of Medicine and Dentistry, Queen Mary University of London. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


1. Jurkowska RZ, Jurkowski TP, Jeltsch A. Structure and function of mammalian DNA methyltransferases. Chembiochem. 2011;12:206–222. [PubMed]
2. Tsai HC, Baylin SB. Cancer epigenetics: linking basic biology to clinical medicine. Cell Research. 2011;21:502–517. [PMC free article] [PubMed]
3. Baylin SB, Jones PA. A decade of exploring the cancer epigenome - biological and translational implications. Nature Reviews Cancer. 2011;11:726–734. [PMC free article] [PubMed]
4. Yamada Y, Watanabe A. Epigenetic codes in stem cells and cancer stem cells. Adv Genet. 2010;70:177–199. [PubMed]
5. Teh MT, Wong ST, Neill GW, Ghali LR, Philpott MP, et al. FOXM1 is a downstream target of Gli1 in basal cell carcinomas. Cancer Res. 2002;62:4773–4780. [PubMed]
6. Myatt SS, Lam EW. The emerging roles of forkhead box (Fox) proteins in cancer. Nat Rev Cancer. 2007;7:847–859. [PubMed]
7. Koo CY, Muir KW, Lam EW. FOXM1: From cancer initiation to progression and treatment. Biochimica et Biophysica Acta 2011 [PubMed]
8. Gemenetzidis E, Bose A, Riaz AM, Chaplin T, Young BD, et al. FOXM1 upregulation is an early event in human squamous cell carcinoma and it is enhanced by nicotine during malignant transformation. PLoS ONE. 2009;4:e4849. [PMC free article] [PubMed]
9. Waseem A, Ali M, Odell EW, Fortune F, Teh MT. Downstream targets of FOXM1: CEP55 and HELLS are cancer progression markers of head and neck squamous cell carcinoma. Oral Oncol. 2010;46:536–542. [PubMed]
10. Gemenetzidis E, Costea DE, Parkinson EK, Waseem A, Wan H, et al. Induction of human epithelial stem/progenitor expansion by FOXM1. Cancer Res. 2010;70:9515–9526. [PMC free article] [PubMed]
11. Xie Z, Tan G, Ding M, Dong D, Chen T, et al. Foxm1 transcription factor is required for maintenance of pluripotency of P19 embryonal carcinoma cells. Nucleic Acids Res. 2010;38:8027–8038. [PMC free article] [PubMed]
12. Wang Z, Park HJ, Carr JR, Chen Y-J, Zheng Y, et al. FoxM1 in Tumorigenicity of the Neuroblastoma cells and Renewal of the Neural Progenitors. Cancer Research. 2011 DOI: 10.1158/0008-5472.can-1110-4087. [PMC free article] [PubMed]
13. Assou S, Le Carrour T, Tondeur S, Strom S, Gabelle A, et al. A meta-analysis of human embryonic stem cells transcriptome integrated into a web-based expression atlas. Stem Cells. 2007;25:961–973. [PMC free article] [PubMed]
14. Dennis K, Fan T, Geiman T, Yan Q, Muegge K. Lsh, a member of the SNF2 family, is required for genome-wide methylation. Genes Dev. 2001;15:2940–2944. [PMC free article] [PubMed]
15. Teh MT, Gemenetzidis E, Chaplin T, Young BD, Philpott MP. Upregulation of FOXM1 induces genomic instability in human epidermal keratinocytes. Mol Cancer. 2010;9:45. [PMC free article] [PubMed]
16. Dickson MA, Hahn WC, Ino Y, Ronfard V, Wu JY, et al. Human keratinocytes that express hTERT and also bypass a p16(INK4a)- enforced mechanism that limits life span become immortal yet retain normal growth and differentiation characteristics. Mol Cell Biol. 2000;20:1436–1447. [PMC free article] [PubMed]
17. Rheinwald JG, Hahn WC, Ramsey MR, Wu JY, Guo Z, et al. A two-stage, p16(INK4A)- and p53-dependent keratinocyte senescence mechanism that limits replicative potential independent of telomere status. Mol Cell Biol. 2002;22:5157–5172. [PMC free article] [PubMed]
18. Chang SE, Foster S, Betts D, Marnock WE. DOK, a cell line established from human dysplastic oral mucosa, shows a partially transformed non-malignant phenotype. Int J Cancer. 1992;52:896–902. [PubMed]
19. Gaballah K, Costea DE, Hills A, Gollin SM, Harrison P, et al. Tissue engineering of oral dysplasia. J Pathol. 2008;215:280–289. [PubMed]
20. Rheinwald JG, Beckett MA. Tumorigenic keratinocyte lines requiring anchorage and fibroblast support cultures from human squamous cell carcinomas. Cancer Res. 1981;41:1657–1663. [PubMed]
21. Reiss M, Pitman SW, Sartorelli AC. Modulation of the terminal differentiation of human squamous carcinoma cells in vitro by all-trans-retinoic acid. J Natl Cancer Inst. 1985;74:1015–1023. [PubMed]
22. Locke M, Heywood M, Fawell S, Mackenzie IC. Retention of intrinsic stem cell hierarchies in carcinoma-derived cell lines. Cancer Res. 2005;65:8944–8950. [PubMed]
23. Kulkarni PS, Sundqvist K, Betsholtz C, Hoglund P, Wiman KG, et al. Characterization of human buccal epithelial cells transfected with the simian virus 40 T-antigen gene. Carcinogenesis. 1995;16:2515–2521. [PubMed]
24. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55:611–622. [PubMed]
25. Palmisano WA, Divine KK, Saccomanno G, Gilliland FD, Baylin SB, et al. Predicting lung cancer by detecting aberrant promoter methylation in sputum. Cancer Research. 2000;60:5954–5958. [PubMed]
26. Georgiou E, Valeri R, Tzimagiorgis G, Anzel J, Krikelis D, et al. Aberrant p16 promoter methylation among Greek lung cancer patients and smokers: correlation with smoking. European Journal of Cancer Prevention. 2007;16:396–402. [PubMed]
27. Ron-Bigger S, Bar-Nur O, Isaac S, Bocker M, Lyko F, et al. Aberrant epigenetic silencing of tumor suppressor genes is reversed by direct reprogramming. Stem Cells. 2010;28:1349–1354. [PubMed]
28. Bignell GR, Greenman CD, Davies H, Butler AP, Edkins S, et al. Signatures of mutation and selection in the cancer genome. Nature. 2010;463:893–898. [PMC free article] [PubMed]
29. Wonsey DR, Follettie MT. Loss of the forkhead transcription factor FoxM1 causes centrosome amplification and mitotic catastrophe. Cancer Res. 2005;65:5181–5189. [PubMed]
30. Li SK, Smith DK, Leung WY, Cheung AM, Lam EW, et al. FoxM1c counteracts oxidative stress-induced senescence and stimulates Bmi-1 expression. J Biol Chem. 2008;283:16545–16553. [PMC free article] [PubMed]
31. Myant K, Stancheva I. LSH cooperates with DNA methyltransferases to repress transcription. Mol Cell Biol. 2008;28:215–226. [PMC free article] [PubMed]
32. Zhou R, Han L, Li G, Tong T. Senescence delay and repression of p16INK4a by Lsh via recruitment of histone deacetylases in human diploid fibroblasts. Nucleic Acids Res. 2009;37:5183–5196. [PMC free article] [PubMed]
33. Angiero F, Berenzi A, Benetti A, Rossi E, Del Sordo R, et al. Expression of p16, p53 and Ki-67 proteins in the progression of epithelial dysplasia of the oral cavity. Anticancer Research. 2008;28:2535–2539. [PubMed]
34. Itahana K, Zou Y, Itahana Y, Martinez JL, Beausejour C, et al. Control of the replicative life span of human fibroblasts by p16 and the polycomb protein Bmi-1. Molecular and Cellular Biology. 2003;23:389–401. [PMC free article] [PubMed]
35. Kang MK, Kim RH, Kim SJ, Yip FK, Shin KH, et al. Elevated Bmi-1 expression is associated with dysplastic cell transformation during oral carcinogenesis and is required for cancer cell replication and survival. British Journal of Cancer. 2007;96:126–133. [PMC free article] [PubMed]
36. Agger K, Cloos PA, Rudkjaer L, Williams K, Andersen G, et al. The H3K27me3 demethylase JMJD3 contributes to the activation of the INK4A-ARF locus in response to oncogene- and stress-induced senescence. Genes Dev. 2009;23:1171–1176. [PMC free article] [PubMed]
37. Keyes WM, Pecoraro M, Aranda V, Vernersson-Lindahl E, Li W, et al. DeltaNp63alpha Is an Oncogene that Targets Chromatin Remodeler Lsh to Drive Skin Stem Cell Proliferation and Tumorigenesis. Cell Stem Cell. 2011;8:164–176. [PMC free article] [PubMed]
38. Laoukili J, Kooistra MR, Bras A, Kauw J, Kerkhoven RM, et al. FoxM1 is required for execution of the mitotic programme and chromosome stability. Nat Cell Biol. 2005;7:126–136. [PubMed]
39. Timmermann S, Hinds PW, Munger K. Re-expression of endogenous p16ink4a in oral squamous cell carcinoma lines by 5-aza-2′-deoxycytidine treatment induces a senescence-like state. Oncogene. 1998;17:3445–3453. [PubMed]
40. Yen CL, Stone SJ, Cases S, Zhou P, Farese RV., Jr Identification of a gene encoding MGAT1, a monoacylglycerol acyltransferase. Proceedings of the National Academy of Sciences of the United States of America. 2002;99:8512–8517. [PMC free article] [PubMed]
41. Schilling B, Aggeler R, Schulenberg B, Murray J, Row RH, et al. Mass spectrometric identification of a novel phosphorylation site in subunit NDUFA10 of bovine mitochondrial complex I. FEBS Letters. 2005;579:2485–2490. [PubMed]
42. Sweeney KJ, Clark GD, Prokscha A, Dobyns WB, Eichele G. Lissencephaly associated mutations suggest a requirement for the PAFAH1B heterotrimeric complex in brain development. Mechanisms of Development. 2000;92:263–271. [PubMed]
43. Yan W, Assadi AH, Wynshaw-Boris A, Eichele G, Matzuk MM, et al. Previously uncharacterized roles of platelet-activating factor acetylhydrolase 1b complex in mouse spermatogenesis. Proceedings of the National Academy of Sciences of the United States of America. 2003;100:7189–7194. [PMC free article] [PubMed]
44. Flanagan LA, Chou J, Falet H, Neujahr R, Hartwig JH, et al. Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells. Journal of Cell Biology. 2001;155:511–517. [PMC free article] [PubMed]
45. Vadlamudi RK, Li F, Adam L, Nguyen D, Ohta Y, et al. Filamin is essential in actin cytoskeletal assembly mediated by p21-activated kinase 1. Nature Cell Biology. 2002;4:681–690. [PubMed]
46. Li C, Yu S, Nakamura F, Pentikainen OT, Singh N, et al. Pro-prion binds filamin A, facilitating its interaction with integrin beta1, and contributes to melanomagenesis. Journal of Biological Chemistry. 2010;285:30328–30339. [PMC free article] [PubMed]
47. Wang Y, Kreisberg JI, Bedolla RG, Mikhailova M, deVere White RW, et al. A 90 kDa fragment of filamin A promotes Casodex-induced growth inhibition in Casodex-resistant androgen receptor positive C4-2 prostate cancer cells. Oncogene. 2007;26:6061–6070. [PubMed]
48. Bedolla RG, Wang Y, Asuncion A, Chamie K, Siddiqui S, et al. Nuclear versus cytoplasmic localization of filamin A in prostate cancer: immunohistochemical correlation with metastases. Clinical Cancer Research. 2009;15:788–796. [PMC free article] [PubMed]
49. Castoria G, D'Amato L, Ciociola A, Giovannelli P, Giraldi T, et al. Androgen-induced cell migration: role of androgen receptor/filamin A association. PLoS ONE. 2011;6:e17218. [PMC free article] [PubMed]
50. Zhong Z, Yeow WS, Zou C, Wassell R, Wang C, et al. Cyclin D1/cyclin-dependent kinase 4 interacts with filamin A and affects the migration and invasion potential of breast cancer cells. Cancer Research. 2010;70:2105–2114. [PMC free article] [PubMed]
51. Uramoto H, Akyurek LM, Hanagiri T. A positive relationship between filamin and VEGF in patients with lung cancer. Anticancer Research. 2010;30:3939–3944. [PubMed]
52. Ai J, Huang H, Lv X, Tang Z, Chen M, et al. FLNA and PGK1 are two potential markers for progression in hepatocellular carcinoma. Cellular Physiology and Biochemistry. 2011;27:207–216. [PubMed]
53. Burton ER, Gaffar A, Lee SJ, Adeshuko F, Whitney KD, et al. Downregulation of Filamin A interacting protein 1-like is associated with promoter methylation and induces an invasive phenotype in ovarian cancer. Molecular Cancer Research. 2011;9:1126–1138. [PMC free article] [PubMed]
54. Kitagawa H, Izumikawa T, Uyama T, Sugahara K. Molecular cloning of a chondroitin polymerizing factor that cooperates with chondroitin synthase for chondroitin polymerization. Journal of Biological Chemistry. 2003;278:23666–23671. [PubMed]
55. Kalathas D, Theocharis DA, Bounias D, Kyriakopoulou D, Papageorgakopoulou N, et al. Chondroitin synthases I, II, III and chondroitin sulfate glucuronyltransferase expression in colorectal cancer. Mol Med Report. 2011;4:363–368. [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • BioProject
    BioProject links
  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • Gene (nucleotide)
    Gene (nucleotide)
    Records in Gene identified from shared sequence and PMC links.
  • GEO DataSets
    GEO DataSets
    Gene expression and molecular abundance data reported in the current articles that are also included in the curated Gene Expression Omnibus (GEO) DataSets.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • HomoloGene
    HomoloGene clusters of homologous genes and sequences that cite the current articles. These are references on the Gene and sequence records in the HomoloGene entry.
  • MedGen
    Related information in MedGen
  • Nucleotide
    Primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • Protein
    Protein translation features of primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...