Head and neck squamous cell carcinoma (HNSCC) affects about 700.000 individuals per year worldwide with oral squamous cell carcinoma (OSCC) as a major subcategory. Despite a comprehensive treatment concept including surgery, radiation, and chemotherapy the 5-year survival rate is still only about 50 percent. Chronic inflammation is one of the hallmarks of carcinogenesis. Until now, little is known about the premalignant status of oral lichen planus (OLP) and molecular alterations in OLP are still poorly characterized. Our study aims to delineate differential DNA methylation patterns in OLP, OSCC, and normal oral mucosa.
By applying a bead chip approach, we identified altered chromosomal patterns characteristic for OSCC while finding no recurrent alterations in OLP. In contrast, we identified numerous alterations in the DNA methylation pattern in OLP, as compared to normal controls, that were also present in OSCC. Our data support the hypothesis that OLP is a precursor lesion of OSCC sharing multiple epigenetic alterations with OSCC.
Citation: Németh CG, Röcken C, Siebert R, Wiltfang J, Ammerpohl O, Gassling V (2019) Recurrent chromosomal and epigenetic alterations in oral squamous cell carcinoma and its putative premalignant condition oral lichen planus. PLoS ONE 14(4): e0215055. https://doi.org/10.1371/journal.pone.0215055
Editor: Hiromu Suzuki, Sapporo Ika Daigaku, JAPAN
Received: September 3, 2018; Accepted: March 26, 2019; Published: April 9, 2019
Copyright: © 2019 Németh et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Head and neck squamous cell carcinoma (HNSCC affects about 700,000 individuals per year currently making it the sixth leading cause of cancer-related mortality worldwide [1, 2]. Among these, HNSCC of the lip, oral cavity and pharynx have been estimated to account for 529,500 incident cases and 292,300 deaths in 2012 . Despite a comprehensive treatment concept including surgery, radiation, and chemotherapy the 5-year survival rate of patients carrying these tumours is still only about 50 percent . Thus, further insight into HNSCC pathogenesis is urgently needed to improve preventive and therapeutic strategies.
The development of HNSCC is a multistep process characterised by the accumulation of genetic and epigenetic alterations leading to activation of oncogenes and inactivation or loss of tumour suppressor genes , including frequent DNA copy number gains at chromosomes 3q, 5p and 8q, as well as copy number losses on 3p and 8p . To date, several driver genes in HNSCC like TP53, CDKN2A, PIK3CA, HRAS, and FBXW7 have been identified .
Besides genetic alterations, virtually all cancers are associated with aberrant DNA methylation, in particular, a genome-wide hypomethylation of repetitive sequences and hypermethylation of high CpG-content promoters and target genes of the polycomb repressive complexes (PRC1 and PRC2) [6–15]. This results in genomic instability and contributes to altered gene expression and cell differentiation .
Locus-specific and global alterations of methylation have also been reported in HNSCC [17–20]. Recently, the global hypomethylation and gene-specific methylation processes have been observed in a series of 138 HNSCCs and the authors proposed that clinical characteristics and exposures influencing disease processes lead to HNSCC . Furthermore, it has been shown that hypermethylation among 13 CpG loci, characterized by polycomb gene targets, mammalian interspersed repetitive elements and transcription factor binding sites was associated with reduced survival in patients with HNSCC . A number of locus-specific studies have investigated the methylation status of certain tumour-suppressor genes. In this context, it has been proposed that the inactivation of p16 might play a crucial role in carcinogenesis of several cancers including HNSCC . Concerning the development of oral cancer hypermethylation of p16Ink4a and p14ARF was detected in 57.7% respectively 3.8% of patients with oral epithelial dysplasia . The highest rates of hypermethylation occurred in lesions of the tongue and mouth base. The hypermethylation of promoters of cancer-related genes in HNSCC has also been widely reported [25–28]. The wide range of frequencies observed in different studies was explained by varying environmental factors encountered by each population and distinct genetic pathways involved .
Chronic inflammation is known to be a crucial factor at the interface between genetics and the environment which may contribute to cancer development. Oral lichen planus (OLP), a chronic mucocutaneous autoimmune disease of unknown aetiology is still being discussed as a precursor lesion of OSCC, while controversial results do not allow for definitive conclusions [30–32]. In contrast to premalignant lesions, such as leukoplakia and erythroplakia, proof of molecular alterations in OLP is still lacking [33, 34] and data concerning methylation patterns in OLP are very limited. Recently, the investigation of prognostic biomarkers in OLP and oral squamous cell carcinoma (OSCC) showed that p16[CDKN2A] methylation and miR-137 promoter methylation occur with a frequency of 25% and 35% respectively in patients with OLP, and 50% and 58.3% respectively, in patients with OSCC. In healthy subjects, however, p16[CDKN2A] methylation and miR-137 promoter methylation were virtually absent . A recently published study aimed to assess the methylation status of different candidate genes in OSCC, high-grade squamous intraepithelial lesions (HG-SIL), low-grade SIL (LG-SIL), OLP, and eight healthy donors in order to detect early OSCC and potential precursor lesions. The authors state that the aberrant DNA methylation of GP1BB and ZAP70 represent promising tools for the early detection of OSCC and HG-SIL .
Our study aims to delineate differential DNA methylation patterns in OLP, OSCC and healthy tissue which might shed further light on the pathogenesis of OSCC and its potential precursor lesions.
Materials and methods
Patient recruitment and sample selection were described in detail previously . Briefly, native tissue samples were collected at the Department of Oral and Maxillofacial Surgery, University Hospital of Schleswig-Holstein, Campus Kiel, Germany. Due to the strict inclusion criteria from initially 117 patients, eight with reticular OLP in their past medical history and typical clinical features were finally selected and specimens were taken at regular intervals. Diagnosis of OLP was verified by board-certified surgical pathologists of the Department of Pathology applying the specified histopathological criteria detailed previously . Fifteen samples of OSCC were obtained during surgery and handled as described above to verify tumour diagnosis and to be stored for further analysis. These samples were taken from inside the main tumour mass but outside of obviously necrotic tissue and underwent a histopathological examination to provide the best possible homogeneity. The control group was composed of eighteen normal oral mucosa samples from elective oral surgery whose anamnestic data and histopathological examination excluded any diseases of the oral mucosa (Table 1 and S1 Table).
The study design complied with the Declaration of Helsinki and was approved by the ethics board of the Christian-Albrechts-University of Kiel, Germany (reference number: D 426/08). All patients gave written informed consent before being included in our study.
Human papillomavirus status
All patient samples had to be negative for human papillomavirus (HPV) infection, which was tested as follows. DNA was extracted from formalin fixed and paraffin embedded tissue sections using the QIAamp DNA Mini Kit (Qiagen, Hilden) and then centrifuged for further purification. The isolated DNA was checked for its concentration and quality using the BIOMED-2-protocol and after verification was amplified via PCR. For this the primer sets MY09/11 and GP5+/6+ targeting the HPV-specific L1-region were used, having been tested for reliability and repeatability .
Array-based DNA-methylation analysis
Array-based DNA-methylation analysis using the HumanMethylation450 BeadChip was performed as described in detail previously . DNA was extracted from fresh frozen tissue using standard methods. To prevent a sex-related analysis bias, data of loci located on chromosomes X and Y were excluded. After quality filtering a total of 41 hybridizations, 471,336 loci entered the final analyses (Table 1 and S1 Table). Results are available in a MIAMI compliant format from Gene Expression Omnibus (GSE123781). Data were exported from the Genome Studio software and the Omics Explorer (ver.3.0 (25); Qlucore, Lund, Sweden) was used for further analyses. False discovery rate (FDR) and/or variance (σ/σmax) filters applied for individual analyses are given in the respective descriptions of the results. Microsoft Excel 2010 and R (R Core Team. R: A Language and Environment for Statistical. R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org) were used for graphical presentation of the data.
Analyses of chromosomal alterations
The R-package “CopyNumber450k” applying the standard settings on the HumanMethylation450 BeadChip data was used to analyse chromosomal alterations (https://bioconductor.riken.jp/packages/3.1/bioc/html/CopyNumber450k.html) in OLP, OSCC as well as control tissue samples. The control data set provided with the R-package acted as a reference for identifying genetic alterations (R-package “CopyNumber450kData”). P-values <0.01 as determined by the software package were considered as significant. Furthermore, chi-square statistics to calculate enrichment of particular characteristics as well as further statistic tests have been performed using R (R Core Team. R: A Language and Environment for Statistical. R Foundation for Statistical Computing, Vienna, Austria, https://www.R-project.org) as well as Prism software (ver. 4.02; Graph Pad Software, San Diego, CA).
Analyses of the chromosomal state
Information on NHEK (normal human epidermal keratinocytes) cells provided by the Ensembl database (http://www.ensembl.org/info/website/news_by_topic.html?db=core; topic = regulation) was used to correlate each individual locus present on the HumanMethylation450k BeadChip with a unique chromosomal state in this cell type. Enrichment of aberrantly methylated loci in OLP or OSCC as compared to normal controls located in distinct regulatory regions was calculated. A list of all loci included in the differential methylation analysis acted as a reference list. Enrichments of loci located in CpG-islands or known differentially methylated regions (DMR) were calculated based on the information provided by Illumina with a list of all CpG-loci included into the study acting as background/reference list by performing a chi-square statistic.
Calculation of methylation age
To calculate the epigenetic age of samples based on the DNA methylation pattern we used the DNAm age calculator provided at https://dnamage.genetics.ucla.edu/
Gene ontology analysis
Gene ontology analysis for enrichment of biological processes, molecular function and protein binding among differentially methylated genes was performed using the Gene Ontology enRIchmentanaLysis and visuaLizAtion tool GOrilla, http://cbl-gorilla.cs.technion.ac.il/, accessed 01/2016 [40, 41].
Recurrent chromosomal alterations in oral lichen planus and oral squamous cell carcinoma
In the first step, we analysed the number as well as the size of genetic copy number variations (CNV) in control tissue, OLP and OSCC for sequence gains and losses separately. For this, we applied the approach described by Feber and colleagues, who reported suitability of HumanMethylation450k BeadChip data for CNV analysis . While no significant difference in either the number of CNVs between controls and OLPs or the sizes of gained or lost fragments could be detected, the number of CNVs observed in OSCC was significantly increased as compared to controls (p<0.01) and OLPs (p<0.01). Furthermore, the size of either gained (control vs. OSCC: p<0.00001; OLP vs. OSCC: p<0.001) or lost regions (control vs. OSCC: p<0.0001; OLP vs. OSCC p<0.01) was significantly increased in OSCC as compared to controls or OLPs (Fig 1).
The absolute number (counts) of detected alterations in the individual groups (A), as well as the size of losses (B) and gains (C), are presented as box plots indicating quartiles and medians.
Additionally, we wondered whether we could identify recurrent copy number variations in our data on OSCC or OLP samples as compared to the corresponding normal control samples. While we did not detect significant recurrent alterations in any of the OLP samples, we found recurrent genetic alterations in all OSCC samples. All analysed OSCCs displayed at least a gain of chromosomal material at chromosome 9 or a combination of gaining a region of chromosome 8 together with the simultaneous loss of genetic information at chromosome 3 (Fig 2 and S1A–S1C Fig).
Detailed presentation of chromosomal alterations in normal oral squamous cells (A), oral lichen planus (B) and oral squamous cell carcinoma (C, D). For this presentation, four typical examples have been selected. The numbers below the diagrams indicate the chromosomes, green: significant chromosomal gains, red: significant chromosomal losses (as determined by CopyNumber450k).
Seven of the 15 cancer specimens showed a gain of genomic material in chromosome 9. Interestingly, in these seven samples (K13, K17, K23, K35, K39, K43, and K45) the alteration started at position 11,410 (as detected by the array) and ended between 6,681,608 and 141,111,396. This restricts the minimal fragment length to 6,670,198bp. The affected genomic region contains at least 39 genes, including e.g. KANK1, DMRT1, DMRT3, UHRF2, JAK2, SMARCA2, SLC1A1, IL33, ERMP1, and FOXD4 (S2 Table). None of the control or OLP samples showed this gain.
Additionally, 10 of 15 OSCCs (K13, K16, K17, K20, K23, K25, K38, K39, K43, and K45), as well as one OLP (L1), showed losses at chromosome 3 starting at position 114,243 (as determined by the array). The end position ranged between 427,757 and 122,296,369. The minimal shared region contains the gene coding for CHL1 (cell adhesion molecule L1 like). Several additional genomic regions at chromosome 3 and 8 showed losses commonly detected in several specimens, however, these were detected only in a minority of specimens or also found in control specimens (S2 Table).
Identification of recurrent aberrant methylation at CpG-loci in oral lichen planus and oral squamous cell carcinoma cells
To identify differentially methylated CpG loci in OLP, OSCC, as well as healthy control tissues, an ANOVA analysis on data obtained from 450k BeadChip analysis was performed. Applying an FDR<1x10-9, we identified 2,192 CpG loci corresponding to 978 genes differentially methylated (S3 Table). A principal component analysis (Fig 3A) as well as a hierarchical cluster analysis (Fig 3B) based on these loci separated the normal controls from the OLP and subsequently both groups from the OSCC samples. Interestingly, DNA methylation of these CpG-loci in the distinct sample groups was clearly homogeneous. OLP samples showed an intermediate DNA methylation pattern between controls and carcinoma samples but were more similar to the former than the latter. The intermediate state of OLP between healthy tissue and OSCC is also supported by further analyses applying a t-test statistic to compare controls and OSCC (S3 Fig). In a principal component analysis based on 2,548 loci distinguishing controls and OSCC (FDR<1.08x10-10), OLP samples located mainly between controls and OSCC samples. The same holds true in a PCA of the 7,611 CpG loci showing the highest variance in the data set (σ/σmax≥0.4) as identified by an unsupervised data analysis approach (S4 Fig). Two samples (a single carcinoma and a single control sample) clustered with OLP samples. This might be due to tissue heterogeneity and different compartments within these samples.
Principal component analysis (A) and hierarchical cluster analysis (B) of OSCC (red spheres and boxes), OLP (blue) and normal controls (green) based on differentially methylated loci (FDR<1.0x10-9, ANOVA, n = 2192 loci). Heatmap (B): blue: low DNA methylation, yellow: high DNA methylation. For presentation purposes, mean DNA methylation values were normalized to zero (mean = 0).
Localisation of differentially methylated CpG loci in regions of distinct chromosomal states
To get further insight into the localisation of differentially methylated CpG-loci in chromosomal regions with distinct chromatin states, we used data on the chromosomal state in NHEK cells available from the Ensembl database. While the identified regions containing differentially methylated loci were depleted of active promoters (3.62fold, p<0.0001) as well as heterochromatic regions (2.52fold, p<0.0001) in NHEK cells, we found a significant enrichment of differentially methylated loci in repressed chromatin (2.47fold, p<0.0001), transcription transition (2.00fold, p<0.0001), enhancer (1.5fold, p<0.0001) and in particular in poised promoter (3.18fold, p<0.0001) sequences of NHEK cells (Fig 4). Furthermore, it turned out that the aberrantly methylated loci were also enriched for loci located in CpG-islands, DNase I hypersensitive sites, enhancers and known differentially methylated regions (S2 Fig).
To determine the individual chromosomal state, data provided by the Ensembl database on the chromosomal state in NHEK cells were used. Pie charts of the chromatin states of differentially methylated loci (FDR<1.0x10-9, ANOVA, n = 2192 loci) (A) and all loci present on the HumanMethylation450 BeadChip included into this study (B) are presented. (C) Bar chart showing odds ratios of selected chromosomal states affected by differential DNA methylation in OSCC, OLP, and controls. ***: p<0.001; **: p<0.01; *: p<0.05; ns: not significant.
Gene ontology analysis
A gene ontology analysis of the genes corresponding to the 2,192 different loci using the GOrilla tool revealed 117 significantly enriched processes (FDR<0.01). Of these 39 processes were enriched ≥2 fold (S4 Table). Nine GO terms were enriched ≥4 fold: glandular epithelial cell differentiation (6.79 fold), signal transduction involved in regulation of gene expression (6.54 fold), proximal/distal pattern formation (6.23 fold), forelimb morphogenesis (5.75 fold), embryonic forelimb morphogenesis (5.66 fold), cell differentiation in spinal cord (4.67 fold), cell fate specification (4.52 fold), dorsal/ventral pattern formation (4.22 fold) and peptide hormone secretion (4.00 fold). Overall this list contains processes involved e.g. in cell differentiation, embryonic development, and signal transduction.
Oral lichen planus as a putative precursor lesion of oral squamous cell carcinoma
In a follow-up study with 327 patients, the annual malignant transformation rate amounted to about 0.5%  and OLP has been considered a putative precursor lesion of OSCC. Therefore we tested the hypothesis that genes becoming aberrantly methylated in the (putative) first step, i.e. development of OLP from healthy tissue, and those genes which become aberrantly methylated during the putative second step, i.e. progression of OLP to OSCC, show distinct characteristics. Our approach is similar to the one previously described in hepatocellular carcinoma .
We classified the 2,192 differentially methylated loci identified above into i) those hypermethylated in both OLP and OSCC, as compared to the control samples (n = 196; OLP and OSCC vs. control), ii) loci hypermethylated (n = 634; OLP and control vs. OSCC) or iii) loci hypomethylated (n = 6; OLP and control vs. OSCC) exclusively in OSCC (Fig 5). For this approach, only loci showing a difference of the means (delta.beta) greater than 0.2 between the two groups compared were considered as differentially methylated. 1,356 loci did not meet this criterion and were therefore excluded from this analysis; no locus was found to be hypomethylated in both OSCC and OLP, as compared to controls. One locus (cg02188358, FLT1) showed hypermethylation in OLP as compared to control samples (delta.beta(OLP-control)>0.2), as well as in OSCC as compared to OLP (delta.beta(OSCC-OLP)>0.2)(p<0.001; Fig 6). Loci hypermethylated in OLP and OSCC were significantly depleted of loci located in active promoters (OR: 0.20, p<0.0001) and regions of heterochromatin in NHEK cells (OR: 0.34, p<0.0001). In contrast, this group was enriched for loci located in poised promoters (OR: 2.73, p<0.0001), weak enhancers (OR: 1.89, p<0.05) and repressed chromatin (OR: 2.73, p<0.0001). Similarly, the set of CpG loci hypermethylated exclusively in OSCC was also enriched for loci located in poised promoters (OR: 2.82, p<0.0001), weak enhancers (OR: 1.56, q<0.05) as well as repressed genomic sequences (OR: 2.51, p<0.0001) in NHEK cells, whereas it was depleted for loci located in active promoters (OR: 0.29, p<0.0001) and heterochromatic regions (OR: 0.42, p<0.0001).
Methylation gained from normal control to OLP (A), loci exclusively hypermethylated in oral squamous cell carcinoma (OSCC) (gaining methylation in OSCC; B) and loci continuously losing DNA methylation from controls via OLP to OSCC (C). Different colors of lines indicate individual CpG loci.
Epigenetic age of oral lichen planus and oral squamous cell carcinoma samples
Because aging is correlated with distinct changes in the DNA methylome, DNA methylation can be used to calculate the (biological) age of a sample  and deviations of the epigenetic age from the chronological age have been associated with numerous diseases, including several cancer entities . Therefore we used the DNAm age (DNA Methylation Age) calculator to determine the epigenetic age of the OSCC, OLP, as well as control samples (S5 Fig). However, we found no marked acceleration or deceleration of aging, neither in normal oral tissue samples nor in OLP or OSCC.
By assaying more than 450,000 CpG sites in the genome, we studied the patterns of altered methylation between normal oral mucosa, OLP, and OSCC. In this analysis, distinct patterns of altered methylation arise that may contribute to a better understanding of OSCC and its carcinogenesis.
The aetiology of OLP and its potential to represent a premalignant phenotype remain poorly understood [30, 32, 47]. However, there is strong evidence concerning the relationship between inflammation and tumour development. It is obvious that the oral cavity with its special environmental conditions, e.g. bacterial and viral infections and continuing chemical irritations, is of particular risk to develop chronic inflammatory stress. It is known that the interplay of inflammatory microenvironment and genetic and epigenetic alterations can lead to the development of various types of cancer over time [48, 49]. Moreover, it has been indicated that epigenetic changes occurring in cancer, e.g. increased DNA methylation and gene silencing are accelerated in inflamed tissue .
This study aimed to identify genetic and epigenetic changes found in both OLP and in OSCC, as well as alterations unique to OSCC. This information might help us in defining molecular steps in the development of OLP and the subsequent onset of OSCC . While we identified several genetic alterations in OLP none of these were restricted to OLP as they could also be found in normal controls. Only two individual OLP samples showed similarities in their alteration pattern. Thus, at least in the limited number of OLP samples included in this study, no genetic alteration specific for OLP, which might allow differentiation from normal tissue samples, could be identified. Furthermore, the groups of OLP, OSCC and control patients differed in their median age. The healthy control group consisted mainly of rather young people who underwent elective wisdom tooth removal or orthognathic surgery; the development of OLP and especially OSCC occurs at a more advanced age. A greater age means longer exposure to environmental factors and possibly a different lifestyle, which could influence methylation patterns and accelerate cancer risk . For these reasons a larger study with matched groups could reduce related confounders. In contrast, genetic alterations characteristic for OSCC have been found on the chromosomes 3, 8 and 9. Sequences recurrently gained on chromosome 9 include the genes encoding the KN motif and ankyrin repeat domains 1 (KANK1), involved in induction of apoptosis [52, 53], as well as known key players of carcinogenesis in gastrointestinal and haematological neoplasia like UHRF2 protein ligase (ubiquitin-like with PHD and ring finger domains 2), JAK2 (Janus kinase 2) or SMARCA2 (SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily a, member 2) [54–59]. While these gained sequences on chromosome 9 were found in 47% of OSCC, none of the control or OLP samples carried them. Furthermore, 67% of all OSCCs showed sequence losses at chromosome 3, starting at position 114,243. The minimal region lost in these samples carried only the gene encoding for CHL1 (cell adhesion molecule L1 like). The encoded protein may be involved in signal transduction pathways. Alterations of CHL1 have been correlated with several tumour entities and have been suggested to function as tumour biomarkers e.g. in breast cancer [60–63]. Interestingly, only one OLP case also showed this specific alteration, which might possibly argue for an advanced premalignant state of this particular sample. Sequence gains were also identified on chromosome 8 in several patients; however, these were not highly recurrent. Overall, we identified recurrent genetic alterations in OSCC but not in OLP. Number and size of the genetic alterations found in OSCC were significantly increased as compared to OLP and normal controls, arguing for increased genetic instability in the neoplastic samples. Since no change in number and size of genetic alteration could be detected when comparing OLP and controls, our data argue that genetic stability is not significantly impaired in OLP.
In contrast to the genetic alterations, we identified numerous alterations in the DNA methylation pattern of both OLP and OSCC, as compared to the control tissues. DNA methylome data clearly distinguished all three sample groups. Interestingly, a large number of CpG loci aberrantly methylated in OLP were found equally methylated in OSCC. This supports the hypothesis that OLP is a precursor lesion of OSCC; parts of the DNA methylome are already altered and thus may prepare the affected (precancerous) cell to transform into a cancer cell. However, numerous additional epimutations were uniquely characteristic of OSCC. This holds particularly true for the hypermethylation of loci located in bivalent promoters. Interestingly, similar findings have been reported in liver diseases (fatty liver disease, liver cirrhosis and hepatocellular carcinoma (HCC)). Here, increasing numbers of epimutations were reported during disease progression, while hypermethylation of bivalent promoters and polycomb repressor complex 2 (PRC2) target genes was restricted to HCC [44, 64]. In line with our findings, hypermethylation of bivalent promoters has been described as a general characteristic of malignant tumours .
In summary, the results of our study support the hypothesis that OLP is a precursor lesion of OSCC. While OLP and OSCC already share numerous epigenetic alterations, further epimutations, as well as genetic alterations, are characteristic for OSCC. Our study findings might help to improve early diagnosis, to identify patients at risk for the development of OSCC and to ensure appropriate patient management.
S1 Table. Detailed overview of the samples included in the DNA methylation analyses.
S2 Table. Detailed overview of chromosomal alterations of the samples included in the DNA methylation analyses.
S3 Table. Detailed overview of methylated CpG-loci of the samples included in the DNA methylation analyses.
S4 Table. Detailed overview of the 117 significantly enriched processes of the gene ontology analysis.
S1 Fig. Detailed presentation of gains and losses of genomic regions in all individual samples included in this study.
The numbers below of the diagrams indicate the chromosomes, green: significant chromosomal gains, red: significant chromosomal losses (as determined by CopyNumber450k). Data from oral squamous cell carcinoma (A), oral lichen planus (B) and normal controls (C) are presented.
S2 Fig. Enrichment of differentially methylated loci located in distinct chromosomal states.
To determine the individual chromosomal state data provided by the allocation of individual loci present on the HumanMethylation450 BeadChip provided by Illumina was used. The bar chart shows fold enrichment of selected chromosomal states affected by differential DNA methylation in OSCC, OLP, and controls. ***: p<0.001; **: p<0.01; *: p<0.05; ns: not significant.
Hierarchical cluster analysis (A) and principal component analysis (B) based on the DNA methylation values of 2,548 CpG loci differentiating control samples (green spheres and green boxes) and oral squamous cell carcinoma (red spheres and red boxes) (FDR<1.08x10-10, t-test). Heatmap (A): blue: low DNA methylation, yellow: high DNA methylation. For presentation purposes, mean DNA methylation values were normalized to zero (mean = 0). PCA: blue spheres indicate oral lichen planus samples (position of OLP in PCA is shown, OLP data was not included in t-test statistic to determine loci differentially methylated between controls and OSCC).
S4 Fig. Principal component analysis of 7,611 CpG with high variance in the data set (unsupervised approach, σ/σmax≥0.4).
Green spheres: control samples, blue: oral lichen planus (OLP) samples, red: oral squamous cell carcinoma (OSCC).
S5 Fig. Scatterplot demonstrating the correlation of the chronological age of the patient with the epigenetic age (“methylation age”) as determined by Horvath’s age calculator.
Blue dots: control samples, green: oral lichen planus, red: oral squamous cell carcinoma. The black line indicates positions with chronological age = methylation age.
The technical assistance of L. Valles is gratefully acknowledged. We would like to deeply acknowledge all participating patients and healthy volunteers.
- 1. Kademani D. Oral cancer. Mayo Clin Proc. 2007;82(7):878–87. pmid:17605971.
- 2. Ferlay J SI EM, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. GLOBOCAN 2012 v1.0 CancerBase No. 11 [Internet]. Lyon, France: International Agency for Research on Cancer: International Agency for Research on Cancer; 2013 [cited 2016 24/01/2016]. Available from: http://globocan.iarc.fr.
- 3. Mascolo M, Siano M, Ilardi G, Russo D, Merolla F, De Rosa G, et al. Epigenetic disregulation in oral cancer. Int J Mol Sci. 2012;13(2):2331–53. pmid:22408457; PubMed Central PMCID: PMCPMC3292026.
- 4. Cancer Genome Atlas N. Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature. 2015;517(7536):576–82. pmid:25631445; PubMed Central PMCID: PMCPMC4311405.
- 5. Agrawal N, Frederick MJ, Pickering CR, Bettegowda C, Chang K, Li RJ, et al. Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science. 2011;333(6046):1154–7. pmid:21798897; PubMed Central PMCID: PMCPMC3162986.
- 6. Costello JF, Fruhwald MC, Smiraglia DJ, Rush LJ, Robertson GP, Gao X, et al. Aberrant CpG-island methylation has non-random and tumour-type-specific patterns. NatGenet. 2000;24(2):132–8. pmid:10655057
- 7. Robertson KD, Wolffe AP. DNA methylation in health and disease. Nat Rev Genet. 2000;1(1):11–9. pmid:11262868.
- 8. Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet. 2002;3(6):415–28. pmid:12042769.
- 9. Costello JF, Plass C, Cavenee WK. Aberrant methylation of genes in low-grade astrocytomas. Brain Tumor Pathol. 2000;17(2):49–56. pmid:11210171.
- 10. Faust C, Lawson KA, Schork NJ, Thiel B, Magnuson T. The Polycomb-group gene eed is required for normal morphogenetic movements during gastrulation in the mouse embryo. Development. 1998;125(22):4495–506. pmid:9778508.
- 11. Peng JC, Valouev A, Swigut T, Zhang J, Zhao Y, Sidow A, et al. Jarid2/Jumonji coordinates control of PRC2 enzymatic activity and target gene occupancy in pluripotent cells. Cell. 2009;139(7):1290–302. pmid:20064375; PubMed Central PMCID: PMCPMC2911953.
- 12. Bracken AP, Helin K. Polycomb group proteins: navigators of lineage pathways led astray in cancer. Nat Rev Cancer. 2009;9(11):773–84. pmid:19851313.
- 13. Schlesinger Y, Straussman R, Keshet I, Farkash S, Hecht M, Zimmerman J, et al. Polycomb-mediated methylation on Lys27 of histone H3 pre-marks genes for de novo methylation in cancer. Nat Genet. 2007;39(2):232–6. pmid:17200670.
- 14. Gal-Yam EN, Egger G, Iniguez L, Holster H, Einarsson S, Zhang X, et al. Frequent switching of Polycomb repressive marks and DNA hypermethylation in the PC3 prostate cancer cell line. Proc Natl Acad Sci U S A. 2008;105(35):12979–84. pmid:18753622; PubMed Central PMCID: PMCPMC2529074.
- 15. Martin-Subero JI, Kreuz M, Bibikova M, Bentink S, Ammerpohl O, Wickham-Garcia E, et al. New insights into the biology and origin of mature aggressive B-cell lymphomas by combined epigenomic, genomic, and transcriptional profiling. Blood. 2009;113(11):2488–97. pmid:19075189.
- 16. Gasche JA, Hoffmann J, Boland CR, Goel A. Interleukin-6 promotes tumorigenesis by altering DNA methylation in oral cancer cells. Int J Cancer. 2011;129(5):1053–63. pmid:21710491; PubMed Central PMCID: PMCPMC3110561.
- 17. Calmon MF, Colombo J, Carvalho F, Souza FP, Filho JF, Fukuyama EE, et al. Methylation profile of genes CDKN2A (p14 and p16), DAPK1, CDH1, and ADAM23 in head and neck cancer. Cancer GenetCytogenet. 2007;173(1):31–7. S0165-4608(06)00610-8 [pii]; pmid:17284367
- 18. Ha PK, Califano JA. Promoter methylation and inactivation of tumour-suppressor genes in oral squamous-cell carcinoma. Lancet Oncol. 2006;7(1):77–82. S1470-2045(05)70540-4 [pii]; pmid:16389187
- 19. Loyo M, Brait M, Kim MS, Ostrow KL, Jie CC, Chuang AY, et al. A survey of methylated candidate tumor suppressor genes in nasopharyngeal carcinoma. Int J Cancer. 2011;128(6):1393–403. pmid:20473931
- 20. Sharma R, Panda NK, Khullar M. Hypermethylation of carcinogen metabolism genes, CYP1A1, CYP2A13 and GSTM1 genes in head and neck cancer. Oral Dis. 2010;16(7):668–73. pmid:20846153
- 21. Poage GM, Houseman EA, Christensen BC, Butler RA, Avissar-Whiting M, McClean MD, et al. Global hypomethylation identifies Loci targeted for hypermethylation in head and neck cancer. ClinCancer Res. 2011;17(11):3579–89. 1078-0432.CCR-11-0044 [pii]; pmid:21505061
- 22. Poage GM, Butler RA, Houseman EA, McClean MD, Nelson HH, Christensen BC, et al. Identification of an epigenetic profile classifier that is associated with survival in head and neck cancer. Cancer Res. 2012;72(11):2728–37. 0008-5472.CAN-11-4121-T [pii]; pmid:22507853
- 23. Miracca EC, Kowalski LP, Nagai MA. High prevalence of p16 genetic alterations in head and neck tumours. BrJ Cancer. 1999;81(4):677–83. pmid:10574255
- 24. Kresty LA, Mallery SR, Knobloch TJ, Song H, Lloyd M, Casto BC, et al. Alterations of p16(INK4a) and p14(ARF) in patients with severe oral epithelial dysplasia. Cancer Res. 2002;62(18):5295–300. pmid:12234999
- 25. de FC-S , Stur E, Agostini LP, de Podesta JR, de Oliveira JC, Soares MS, et al. Promoter hypermethylation in primary squamous cell carcinoma of the oral cavity and oropharynx: a study of a Brazilian cohort. MolBiolRep. 2012;39(12):10111–9. pmid:22936053
- 26. Kaur J, Demokan S, Tripathi SC, Macha MA, Begum S, Califano JA, et al. Promoter hypermethylation in Indian primary oral squamous cell carcinoma. Int J Cancer. 2010;127(10):2367–73. pmid:20473870
- 27. Taioli E, Ragin C, Wang XH, Chen J, Langevin SM, Brown AR, et al. Recurrence in oral and pharyngeal cancer is associated with quantitative MGMT promoter methylation. BMCCancer. 2009;9:354. 1471-2407-9-354 [pii]; pmid:19807915
- 28. Teh MT, Gemenetzidis E, Patel D, Tariq R, Nadir A, Bahta AW, et al. FOXM1 induces a global methylation signature that mimics the cancer epigenome in head and neck squamous cell carcinoma. PLoSOne. 2012;7(3):e34329. PONE-D-11-23825 [pii]. pmid:22461910
- 29. de Martel C, Ferlay J, Franceschi S, Vignat J, Bray F, Forman D, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol. 2012;13(6):607–15. pmid:22575588.
- 30. Eisenberg E. Oral lichen planus: a benign lesion. JOral MaxillofacSurg. 2000;58(11):1278–85. S0278-2391(00)24584-9 [pii]; pmid:11078140
- 31. Fitzpatrick SG, Hirsch SA, Gordon SC. The malignant transformation of oral lichen planus and oral lichenoid lesions: a systematic review. J Am Dent Assoc. 2014;145(1):45–56. pmid:24379329.
- 32. Silverman S Jr. Oral lichen planus: a potentially premalignant lesion. J Oral Maxillofac Surg. 2000;58(11):1286–8. pmid:11078141.
- 33. Ha PK, Benoit NE, Yochem R, Sciubba J, Zahurak M, Sidransky D, et al. A transcriptional progression model for head and neck cancer. ClinCancer Res. 2003;9(8):3058–64.
- 34. Ha PK, Pilkington TA, Westra WH, Sciubba J, Sidransky D, Califano JA. Progression of microsatellite instability from premalignant lesions to tumors of the head and neck. IntJCancer. 2002;102(6):615–7. pmid:12448003
- 35. Dang J, Bian YQ, Sun J, Chen F, Dong GY, Liu Q, et al. MicroRNA-137 promoter methylation in oral lichen planus and oral squamous cell carcinoma. J Oral PatholMed. 2012. pmid:23121285
- 36. Morandi L, Gissi D, Tarsitano A, Asioli S, Monti V, Del Corso G, et al. DNA methylation analysis by bisulfite next-generation sequencing for early detection of oral squamous cell carcinoma and high-grade squamous intraepithelial lesion from oral brushing. J Craniomaxillofac Surg. 2015;43(8):1494–500. pmid:26302938.
- 37. Gassling V, Hampe J, Acil Y, Braesen JH, Wiltfang J, Hasler R. Disease-associated miRNA-mRNA networks in oral lichen planus. PLoS One. 2013;8(5):e63015. pmid:23723971; PubMed Central PMCID: PMCPMC3664564.
- 38. de Souza MMA, Hartel G, Whiteman DC, Antonsson A. Detection of oral HPV infection—Comparison of two different specimen collection methods and two HPV detection methods. Diagn Microbiol Infect Dis. 2018;90(4):267–71. pmid:29317137.
- 39. Kolarova J, Ammerpohl O, Gutwein J, Welzel M, Baus I, Riepe FG, et al. In vivo investigations of the effect of short- and long-term recombinant growth hormone treatment on DNA-methylation in humans. PLoS One. 2015;10(3):e0120463. pmid:25785847; PubMed Central PMCID: PMCPMC4364725.
- 40. Eden E, Lipson D, Yogev S, Yakhini Z. Discovering motifs in ranked lists of DNA sequences. PLoS Comput Biol. 2007;3(3):e39. pmid:17381235; PubMed Central PMCID: PMCPMC1829477.
- 41. Eden E, Navon R, Steinfeld I, Lipson D, Yakhini Z. GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics. 2009;10:48. pmid:19192299; PubMed Central PMCID: PMCPMC2644678.
- 42. Feber A, Guilhamon P, Lechner M, Fenton T, Wilson GA, Thirlwell C, et al. Using high-density DNA methylation arrays to profile copy number alterations. Genome Biol. 2014;15(2):R30. pmid:24490765; PubMed Central PMCID: PMCPMC4054098.
- 43. Bombeccari GP, Guzzi G, Tettamanti M, Gianni AB, Baj A, Pallotti F, et al. Oral lichen planus and malignant transformation: a longitudinal cohort study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2011;112(3):328–34. pmid:21782485.
- 44. Ammerpohl O, Pratschke J, Schafmayer C, Haake A, Faber W, von Kampen O, et al. Distinct DNA methylation patterns in cirrhotic liver and hepatocellular carcinoma. Int J Cancer. 2012;130(6):1319–28. pmid:21500188.
- 45. Horvath S. DNA methylation age of human tissues and cell types. Genome Biol. 2013;14(10):R115. pmid:24138928; PubMed Central PMCID: PMCPMC4015143.
- 46. Perna L, Zhang Y, Mons U, Holleczek B, Saum KU, Brenner H. Epigenetic age acceleration predicts cancer, cardiovascular, and all-cause mortality in a German case cohort. Clin Epigenetics. 2016;8:64. pmid:27274774; PubMed Central PMCID: PMCPMC4891876.
- 47. Zhang L, Michelsen C, Cheng X, Zeng T, Priddy R, Rosin MP. Molecular analysis of oral lichen planus. A premalignant lesion? Am J Pathol. 1997;151(2):323–7. pmid:9250145; PubMed Central PMCID: PMCPMC1858013.
- 48. Balkwill F, Mantovani A. Inflammation and cancer: back to Virchow? Lancet. 2001;357(9255):539–45. S0140-6736(00)04046-0 [pii]; pmid:11229684
- 49. Mantovani A, Allavena P, Sica A, Balkwill F. Cancer-related inflammation. Nature. 2008;454(7203):436–44. nature07205 [pii]; pmid:18650914
- 50. Brower V. Epigenetics: Unravelling the cancer code. Nature. 2011;471(7339):S12–S3. 471S12a [pii]; pmid:21430711
- 51. Dugue PA, Bassett JK, Joo JE, Jung CH, Ming Wong E, Moreno-Betancur M, et al. DNA methylation-based biological aging and cancer risk and survival: Pooled analysis of seven prospective studies. Int J Cancer. 2018;142(8):1611–9. pmid:29197076.
- 52. Guo X, Fan W, Bian X, Ma D. Upregulation of the Kank1 gene-induced brain glioma apoptosis and blockade of the cell cycle in G0/G1 phase. Int J Oncol. 2014;44(3):797–804. pmid:24399197.
- 53. Luo FY, Xiao S, Liu ZH, Zhang PF, Xiao ZQ, Tang CE. Kank1 reexpression induced by 5-Aza-2'-deoxycytidine suppresses nasopharyngeal carcinoma cell proliferation and promotes apoptosis. Int J Clin Exp Pathol. 2015;8(2):1658–65. pmid:25973051; PubMed Central PMCID: PMCPMC4396205.
- 54. Segedi M, Anderson LN, Espin-Garcia O, Borgida A, Bianco T, Cheng D, et al. BRM polymorphisms, pancreatic cancer risk and survival. Int J Cancer. 2016;139(11):2474–81. pmid:27487558.
- 55. Pasic I, Wong KM, Lee JJ, Espin-Garcia O, Brhane Y, Cheng D, et al. Two BRM promoter polymorphisms predict poor survival in patients with hepatocellular carcinoma. Mol Carcinog. 2018;57(1):106–13. pmid:28892201.
- 56. Chen CC, Chiu CC, Lee KD, Hsu CC, Chen HC, Huang TH, et al. JAK2V617F influences epigenomic changes in myeloproliferative neoplasms. Biochem Biophys Res Commun. 2017;494(3–4):470–6. pmid:29066347.
- 57. Xu Y, Jin J, Xu J, Shao YW, Fan Y. JAK2 variations and functions in lung adenocarcinoma. Tumour Biol. 2017;39(6):1010428317711140. pmid:28639892.
- 58. Mori T, Ikeda DD, Yamaguchi Y, Unoki M, Project N. NIRF/UHRF2 occupies a central position in the cell cycle network and allows coupling with the epigenetic landscape. FEBS Lett. 2012;586(11):1570–83. pmid:22673569.
- 59. Lai M, Liang L, Chen J, Qiu N, Ge S, Ji S, et al. Multidimensional Proteomics Reveals a Role of UHRF2 in the Regulation of Epithelial-Mesenchymal Transition (EMT). Mol Cell Proteomics. 2016;15(7):2263–78. pmid:27114453; PubMed Central PMCID: PMCPMC4937503.
- 60. Martin-Sanchez E, Mendaza S, Ulazia-Garmendia A, Monreal-Santesteban I, Blanco-Luquin I, Cordoba A, et al. CHL1 hypermethylation as a potential biomarker of poor prognosis in breast cancer. Oncotarget. 2017;8(9):15789–801. pmid:28178655; PubMed Central PMCID: PMCPMC5362523.
- 61. He LH, Ma Q, Shi YH, Ge J, Zhao HM, Li SF, et al. CHL1 is involved in human breast tumorigenesis and progression. Biochem Biophys Res Commun. 2013;438(2):433–8. pmid:23906755.
- 62. Yang Z, Xie Q, Hu CL, Jiang Q, Shen HF, Schachner M, et al. CHL1 Is Expressed and Functions as a Malignancy Promoter in Glioma Cells. Front Mol Neurosci. 2017;10:324. pmid:29089868; PubMed Central PMCID: PMCPMC5650976.
- 63. Pezzolo A, Sementa AR, Lerone M, Morini M, Ognibene M, Defferrari R, et al. Constitutional 3p26.3 terminal microdeletion in an adolescent with neuroblastoma. Cancer Biol Ther. 2017;18(5):285–9. pmid:28402723; PubMed Central PMCID: PMCPMC5499752.
- 64. Ahrens M, Ammerpohl O, von Schonfels W, Kolarova J, Bens S, Itzel T, et al. DNA methylation analysis in nonalcoholic fatty liver disease suggests distinct disease-specific and remodeling signatures after bariatric surgery. Cell Metab. 2013;18(2):296–302. pmid:23931760.
- 65. Bernhart SH, Kretzmer H, Holdt LM, Juhling F, Ammerpohl O, Bergmann AK, et al. Changes of bivalent chromatin coincide with increased expression of developmental genes in cancer. Sci Rep. 2016;6:37393. pmid:27876760; PubMed Central PMCID: PMCPMC5120258. http://globocan.iarc.fr/