Introduction

Oncology Letters

1792-1074 1792-1082

D.A. Spandidos

10.3892/ol.2024.14223

OL-27-3-14223

Articles

Promoter DNA methylation patterns in oral, laryngeal and oropharyngeal anatomical regions are associated with tumor differentiation, nodal involvement and survival

Rivera-Peña

Bianca

1 2 3 Folawiyo

Oluwasina

4 Turaga

Nitesh

4 Rodríguez-Benítez

Rosa J.

5 Felici

Marcos E.

6 Aponte-Ortiz

Jaime A.

7 Pirini

Francesca

8 Rodríguez-Torres

Sebastián

9 Vázquez

Roger

1 López

Ricardo

1 Sidransky

David

4 Guerrero-Preston

Rafael

4 10 Báez

Adriana

2 3

1Department of Biology, University of Puerto Rico, San Juan 00925, Puerto Rico 2Department of Pharmacology, University of Puerto Rico School of Medicine, San Juan 00936, Puerto Rico 3Department of Otolaryngology-Head and Neck Surgery, University of Puerto Rico School of Medicine, San Juan 00936, Puerto Rico 4Department of Otolaryngology-Head and Neck Surgery, Johns Hopkins University School of Medicine, Baltimore, MD 21231, USA 5Department of General Social Sciences, Faculty of Social Sciences, University of Puerto Rico, San Juan 00925, Puerto Rico 6Oral Health Division, Puerto Rico Department of Health, San Juan 00927, Puerto Rico 7Department of General Surgery, University of Puerto Rico School of Medicine, San Juan 00936, Puerto Rico 8Biosciences Laboratory, IRCCS Instituto Romagnolo per lo Studio dei Tumori ‘Dino Amadori’, Meldola I-47014, Italy 9Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA 10Department of Research and Development, LifeGene-Biomarks, San Juan 00909, Puerto Rico

Correspondence to: Dr Adriana Báez, Department of Otolaryngology-Head and Neck Surgery, University of Puerto Rico School of Medicine, 960 Américo Miranda Avenue, San Juan 00936, Puerto Rico, E-mail: adriana.baez@upr.edu Dr Rafael Guerrero-Preston, Department of Research and Development, LifeGene-Biomarks, San Juan 00909, Puerto Rico, E-mail: rguerrero@lifegenedna.com

03 2024

08 01 2024

27 3

08042023 23112023

2024

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Differentially methylated regions (DMRs) can be used as head and neck squamous cell carcinoma (HNSCC) diagnostic, prognostic and therapeutic targets in precision medicine workflows. DNA from 21 HNSCC and 10 healthy oral tissue samples was hybridized to a genome-wide tiling array to identify DMRs in a discovery cohort. Downstream analyses identified differences in promoter DNA methylation patterns in oral, laryngeal and oropharyngeal anatomical regions associated with tumor differentiation, nodal involvement and survival. Genome-wide DMR analysis showed 2,565 DMRs common to the three subsites. A total of 738 DMRs were unique to laryngeal cancer (n=7), 889 DMRs were unique to oral cavity cancer (n=10) and 363 DMRs were unique to pharyngeal cancer (n=6). Based on the genome-wide analysis and a Gene Ontology analysis, 10 candidate genes were selected to test for prognostic value and association with clinicopathological features. TIMP3 was associated with tumor differentiation in oral cavity cancer (P=0.039), DAPK1 was associated with nodal involvement in pharyngeal cancer (P=0.017) and PAX1 was associated with tumor differentiation in laryngeal cancer (P=0.040). A total of five candidate genes were selected, DAPK1, CDH1, PAX1, CALCA and TIMP3, for a prevalence study in a larger validation cohort: Oral cavity cancer samples (n=42), pharyngeal cancer tissues (n=25) and laryngeal cancer samples (n=52). PAX1 hypermethylation differed across HNSCC anatomic subsites (P=0.029), and was predominantly detected in laryngeal cancer. Kaplan-Meier survival analysis (P=0.043) and Cox regression analysis of overall survival (P=0.001) showed that DAPK1 methylation is associated with better prognosis in HNSCC. The findings of the present study showed that the HNSCC subsites oral cavity, pharynx and larynx display substantial differences in aberrant DNA methylation patterns, which may serve as prognostic biomarkers and therapeutic targets.

differentially methylated regions precision medicine head and neck squamous cell carcinoma survival prognostic biomarkers epigenome-wide analysis

University of Puerto Rico School of Medicine Department of Surgery Otolaryngology Section

NIH/National Cancer Institute

P20CA91402

U54CA96297

K01CA164092

R44CA254690

R44CA281719

U01CA84986

NIH/National Institute of General Medical Sciences

S06GM8224

NIH/National Institute of Dental and Craniofacial Research Award

RC2DE20957

NIH/National Institute of Minority Health and Disparities

R44MD014911

NIGMS-RISE award

R25 GM061838

NIH/NCRR and NIH/NIMHDD

The present study was supported in part by the University of Puerto Rico School of Medicine Department of Surgery Otolaryngology Section; NIH/National Cancer Institute (grant nos. P20CA91402, U54CA96297, K01CA164092, R44CA254690, R44CA281719 and U01CA84986), the NIH/National Institute of General Medical Sciences (grant no. S06GM8224), the NIH/National Institute of Dental and Craniofacial Research Award (grant no. RC2DE20957) and the NIH/National Institute of Minority Health and Disparities (grant no. R44MD014911). Bianca Rivera Peña's research was supported in part by the NIGMS-RISE award (grant no. R25 GM061838). This research used core facilities supported by NIH/NCRR and NIH/NIMHDD awards.

Introduction

Head and neck squamous cell carcinoma (HNSCC) is the 6th most common malignancy and the 8th cause of cancer death worldwide (1,2). HNSCC includes carcinomas from the oral cavity (OSCC), oropharynx (OPSCC), hypopharynx (HPSCC), larynx (LSCC), the paranasal sinuses, and the major and minor salivary glands. The etiology of HNSCC involves a variety of toxic, environmental and viral agents (3). Tobacco and alcohol exposure are the primary etiological factors for HNSCC (4–6). Oncogenic human papillomavirus (HPV) strains, primarily HPV-16, have been recognized as risk factor for HNSCC, particularly for oropharyngeal cancers (7–10). Men are more frequently diagnosed with HNSCC compared with women, and the incidence of HNSCC has a male-to-female ratio of 3:1 in the US (11). This incidence has been changing as women increasingly expose themselves to HNSCC risk factors, tobacco, alcohol and HPV-infection. Park et al (12) showed that women with HNSCC are at a higher risk of dying of the disease than men diagnosed with HNSCC (HR=1.92; 95% CI, 1.07–3.43). However, HPV-associated HNSCC is more common in men compared with women (13). Patients with HPV⁺ HNSCC have a better prognosis than patients with HPV⁻ HNSCC (14); HPV may have a role in the clinical manifestation of this sex disparity. HNSCC is commonly diagnosed in patients ≥60 years old, however, an increasing number of patients are diagnosed with HNSCC at younger ages (15). Most patients with HNSCC are diagnosed at advanced stages of the disease (III or IV), which leads to a poor prognosis outcome (16). HNSCC treatment is generally multimodal including surgery, rand chemoradiation, yet the overall survival (OS) of patients with HNSCC is relatively low, ~2.5 years after treatment, for all HNSCC sites and stages (17).

In the US, African Americans, Hispanics/Latinos and low-income non-Latino-White individuals are at higher risk of developing HNSCC. In Puerto Rico, the incidence of HNSCC is 2.5 times higher than that in Hispanics/Latinos living in the US. The HNSCC incidence of OSCC and OPSCC is 72% higher in Puerto Rico than among Hispanics/Latinos living in the US. Similarly, the incidence of LSCC in Puerto Rico is 51% higher than that in Hispanics/Latinos living in the US (18). Racial and ethnic health disparities are a serious public health concern due to the HNSCC high mortality and morbidity rates, higher treatment costs and the effect on quality of life. Therefore, discovery of actionable targets for the early detection, diagnosis and prognosis of HNSCC, and for guiding treatment would have an immediate impact on reducing these health disparities.

Epigenetic biomarkers, such as aberrant DNA methylation changes, have been used as molecular classifiers for different cancer types, having a predictive capacity for patient prognosis and treatment response (19). Aberrant changes in DNA methylation such as global DNA hypomethylation and specific promoter DNA hypermethylation have been associated with carcinogenesis (20). It has been proposed that aberrant changes in DNA methylation patterns occur early in the carcinogenic process (21).

Aberrant promoter methylation of tumor suppressor genes (TSGs), for instance, CDH1, DAPK1, CDKN2A and RASSF1A, have been detected in HNSCC that resulted in loss of expression and pathway deregulation (22–24). Several studies have demonstrated DNA methylation cancer-related signatures (25,26), suggesting the likelihood of differential DNA methylation patterns among HNSCC anatomical subsites (27). Using a candidate gene approach, the prevalence of the aberrantly methylated TSGs CDKN2A, p14^ARF and CDKN2B in HNSCC tumors was previously evaluated. Bernabe (28) detected aberrant methylation of the TSGs CDKN2A and CDKN2B in HNSCC tumors. A reduction of CDKN2A expression in HNSCC tumors exhibiting methylated (M) CDKN2A was detected with mRNA expression analysis (28). Subsequently, the aberrant methylation of CDH1 was evaluated in HNSCC tumors confirming its occurrence, but hyperM CDH1 was predominantly detected in the larynx compared with other HNSCC subsites (29). Preliminary data suggest that a distinct pattern of aberrant DNA methylation changes may occur in HNSCC anatomic subsites associated with HNSCC heterogeneity and its diverse clinical manifestations.

The primary objective of the present study was to perform a genome-wide DNA methylation analysis in HNSCC samples from three anatomic subsites, oral cavity, oropharynx and larynx, to identify potential DNA methylation targets with prognostic value for HNSCC. Furthermore, a prevalence assessment of selected candidate genes was performed, and their prognostic value was evaluated in an independent HNSCC cohort. It was hypothesized that a biomarker profile based on aberrant DNA methylation specific to every anatomical site, may help clinicians to better diagnose HNSCC, thus providing a more accurate prognosis and identify targets for novel treatments.

Materials and methods <sec> <title>HNSCC discovery cohort

Demographics and clinicopathological characteristics of the HNSCC discovery and prevalence cohorts are shown in Table I. The HNSCC discovery cohort included 21 HNSCC tissue samples from Puerto Rican patients, including 10 OSCC, four OPSCC and seven LSCC samples. The HNSCC discovery cohort samples were compared with 10 healthy oral tissue samples. The mean age of the discovery cohort was 56.62 years (range, 24–76 years; SD, 12.62), and 90 and 10% of patients were male and female, respectively. The HNSCC anatomical subsite distribution included 48, 19 and 33% oral cavity, pharynx and larynx, respectively. Most of the patients with HNSCC were at advanced stages (III/IV; 67%) of the disease. A total of 1/3 of the patients (33%) were HPV⁺. Most tumors showed moderate differentiation (71%). Most samples were obtained from heavy smokers (95%) and heavy drinkers (86%).

HNSCC prevalence cohort

The HNSCC prevalence cohort included 119 HNSCC tissue samples from three anatomical subsites: Oral cavity (n=42), pharynx (n=25) and larynx (n=52). The HNSCC tissue samples of the prevalence cohort were compared with seven healthy oral tissue samples. The mean age of the HNSCC prevalence cohort was 61.2 years (range, 24–98 years; SD, 12.6), and 89.9 and 10.1% were male and female, respectively. The distribution of the HNSCC anatomical subsites included 35.3, 21.0 and 43.7% oral cavity, pharynx and larynx, respectively. Most of the patients with HNSCC were at advanced stages (III/IV; 77%) of the disease. Regarding HPV infection, 47.9% of patients were HPV⁺. Most tumors showed moderate differentiation (65.5%). Most samples were obtained from heavy smokers (87%) and heavy drinkers (84%).

The HNSCC tissue samples for both discovery and prevalence cohorts were obtained from Puerto Rican patients with HNSCC presenting at the School of Medicine Head and Neck Cancer Clinic at the Puerto Rico Medical Center, a tertiary teaching medical center. Patients that were diagnosed with HNSCC through tissue biopsy, and whose tumors were surgically removed signed an informed consent. The tumor tissue collected for the study was analyzed for quality by a pathologist. Oral mucosa samples were obtained from healthy Puerto Rican patients undergoing a routine tooth extraction at the School of Dental Medicine, Department of Surgery after having signed an informed consent. All procedures had the approval of the University of Puerto Rico, Medical Sciences Campus Institutional Review Board (IRB; approval no. MSC-IRB protocol 2770103), and the Johns Hopkins School of Medicine IRB (approval no. NA_00020633). The medical information of the patients with HNSCC was obtained from medical records and pathological reports, including date of diagnosis, site of the primary tumor, tumor grade, date and site of tumor recurrence (if applicable), and date and cause of death. The treatment of choice was surgery followed in some cases by postoperative radiation or chemoradiation. Follow-up information was prospectively collected from either medical records or the Puerto Rican Cancer Registry. Fig. 1 shows an integrated diagram describing the experimental study design.

DNA extraction

Genomic DNA was isolated from all HNSCC and healthy tissues using the DNA Isolation kit for cells and tissues (catalog no. 11814770001; Roche Diagnostics, Ltd.) following the manufacturer's instructions. DNA concentration and quality were measured with the NanoDrop 8000 UV–Vis Spectrophotometer (Thermo Fisher Scientific, Inc.). DNA sample preparation and hybridization to oligonucleotide arrays was carried out at the Head and Neck Cancer Research laboratory, Johns Hopkins School of Medicine.

Detection of HPV-16

Genomic DNA from all HNSCC samples was analyzed for HPV-16 infection. The HPV-16 status was previously detected by immunohistochemistry, end-point PCR and a TaqMan-based quantitative (q)PCR assay, targeting HPV-16 E6 and E7 viral oncogenes. All the HNSCC samples that were classified as HPV-16⁺ had amplification of E6 and E7 viral oncogenes detected through a qPCR assay. HPV-16 E6 and E7 specific primer and probe sets, and qPCR and thermal cycling conditions were previously described (10).

Genome-wide DNA methylation analysis DNA sonication

Two different genomic DNA amounts from HNSCC and healthy samples from the discovery cohort (0.5 and 1 µg) were used as input DNA for sonication to generate 200-800-bp long DNA fragments. DNA sonication was performed in a Covaris E220 ultrasonicator (Covaris, LLC), and analysis of sonicated DNA was performed on the BioAnalyzer 2100 (Agilent Technologies, Inc.) with an Agilent High Sensitivity DNA Kit (catalog no. 5067-4626; Agilent Technologies, Inc.) to verify DNA concentration, quality and purity.

Methylated DNA immunoprecipitation

DNA from HNSCC and healthy samples from the discovery cohort was subjected to methylated DNA immunoprecipitation (MeDIP) using the MagMeDIP kit (Diagenode SA) following the manufacturer's instructions. Two different starting DNA quantities were used (0.5 and 1 µg) for every sonicated sample. A total of 10% of every sample was transferred to a 1.5-ml tube (input DNA) and used as control of the starting DNA material. The remaining 90% of the sonicated DNA was subjected to MeDIP and labeled as immunoprecipitated DNA (IP DNA). IP DNA samples were exposed to a 5-methylcytosine antibody, which recognizes methylated cytosines in the DNA to enrich every sample with methylated DNA. Every tumor or control (healthy) DNA sample had an IP DNA and an input DNA. Samples were subjected to a qPCR assay to determine the efficiency of the MeDIP assay efficiency in enriching methylated DNA. IP DNA samples were compared with input DNA samples to determine if enrichment of methylated DNA occurred. Both DNA samples were tested using four primer pairs included in the MagMeDIP kit (Diagenode SA; Table SI).

The qPCR master mix included the following: 6.25 µl SYBR Green Supermix, 0.5 µl primer pair (10 µM), 3.5 ng either IP DNA or input DNA, and 3.25 µl water. The final reaction volume was 12.5 µl. The thermocycling conditions involved a denaturation step at 95°C for 7 min, followed by 40 cycles at 95°C for 15 sec and at 60°C for 1 min, an incubation step for 95°C for 1 min to denature the DNA, and a melting curve analysis as established by the manufacturer instructions. The efficiency of MeDIP enrichment was calculated using the following equation: % (meDNA-IP/total input)=2^{[(Cq(10%input)-3.32)-Cq(meDNA-IP)] ×100}. The MeDIP recovery was % (meDNA-IP/total input). Samples that showed >50% DNA methylation enrichment were subjected to hybridization and scanning into the 3×720K CpG Island Plus RefSeq Promoter Array (Roche Diagnostics, Ltd.; Fig. S1).

DNA labeling and hybridization

After the MeDIP assay, all DNA samples (IP DNA and input DNA) were subjected to a genome-wide amplification (WGA) assay (Sigma-Aldrich; Merch KGaA) to increase the amount of DNA in every sample. After the WGA assay, DNA samples were purified using the QIAquick PCR Purification Kit (Qiagen Sciences, Inc.). DNA concentration was measured with the NanoDrop 8000 UV–Vis Spectrophotometer (Thermo Fisher Scientific, Inc.). Every DNA sample (IP DNA and input DNA) was labeled with fluorophores using the NimbleGen Dual-Color DNA Labeling Kits (Roche Diagnostics, Ltd.). IP DNA was labeled with Cy5 fluorophore, and the input DNA was labeled with the Cy3 fluorophore. Labeled IP DNA and input DNA samples were combined and hybridized into the 3×720K CpG Island Plus RefSeq Promoter Array. Hybridization was accomplished using the NimbleGen Hybridization Kit (Roche Diagnostics, Ltd.) following standard operating protocol. The 3×720K CpG Island Plus RefSeq Promoter Array allowed hybridization of three samples simultaneously and covered 27,728 annotated CpG islands, 22,532 RefSeq gene promoters, and regulatory elements from the HG18 build. Each promoter array included several positive, negative and non-CpG control regions to calculate experimental performance. Analysis of RefSeq gene promoters involved regions 2.4 kb upstream of the transcription start site (TSS) and 0.6 kb downstream of the TSS for overall coverage of 3 kb of each promoter per gene. Each array was scanned in the NimbleGen MS2 Microarray Scanner (Roche Diagnostics, Ltd.) following the manufacturer's protocol.

Differential methylation bioinformatics

NimbleGen's DeVa's software (Roche Diagnostics, Ltd.) was used to create .xys files from the array's scanned images. The images allowed a peak discovery algorithm to generate an initial list of differentially methylated regions (DMRs) when tumor samples were compared with control samples. The .xys files were used as input data for the analysis using Comprehensive high-throughput arrays for relative methylation (CHARM) bioinformatics package within the R 4.1.2 statistical programming (30). CHARM software is a bioinformatics package used to discover DMRs between samples, calculate the percentage of methylation, verify array quality and control for batch effects. Besides, DMRs were identified with Bump Hunting (version 1.44.0) (31), a statistical genomics tool to identify differential peaks in methylation data. Methylation Bump Hunting is a data analysis pipeline that effectively models measurement error, removes batch effects, detects regions of interest, and attaches statistical uncertainty to regions identified as differentially methylated (32). The bioinformatical analysis pipeline used in the present study included analysis of TSS and CpG Islands independently among tumor and control samples (Data S1). Frequency of genes was analyzed for tumor and control samples. Genes with a frequency of ≥20% in tumor samples were selected. Likewise, commonly occurring genes between tumor and control samples were analyzed. A detailed bioinformatical pipeline description for peak discovery algorithm can be found in Data S1. In summary, the raw intensity data from the array were analyzed and the data was transformed into a log ratio of the intensities of methylated probes vs. unmethylated probes, which represents the M-value. An M-value ~0 represented a similar intensity between the methylated and unmethylated probes. Positive M-values implied that more molecules within the tested probe were methylated than negative M-values, which represented less methylation (32). An M-value cut-off was established to define which CpG targeted sites were aberrantly methylated in tumor samples compared with control samples. CpG targeted sites with an M-value ≥2.0 were classified as methylated and were further analyzed. CpG sites with an M-value <2.0 were classified as unmethylated. CpG targeted sites were also subjected to a low-stringency P-value threshold (P<0.05) and ranked by fold-change between tumor and control samples. A list of DMRs was created using the CpG sites methylation level for every HNSCC subsite. These lists determined the regions in the genome that were differentially methylated between HNSCC and normal samples.

DMR validation in TCGA

DMRs identified through bioinformatical analyses were cross-referenced with available methylation-related databases, including publicly available HNSCC TCGA methylation database and peer-reviewed accessible databases, as previously described (33). Briefly, the Bump Hunting method was used to perform an epigenome-wide analysis of the HNSCC methylome to identify DMRs of biological interest using methylation arrays. Two separate epigenome-wide analyses were carried out using Bioconductor's minfi package (version 1.48.0), as previously described (34). Briefly, an unbiased epigenome-wide DNA methylation analysis was performed using the minfi package in Bioconductor to identify DMRs in 274 primary chemotherapy-naïve HNSCC samples (TCGA dataset) and 32 frequency-matched uvulopalatopharyngoplasty (UPPP) controls (Johns Hopkins Head and Neck Cancer Research laboratory). The significant DMRs (P<0.001) were identified in a CpG island located ≤200 bp upstream and downstream from the 5′ end of the gene. The DMRs results obtained with MeDIP were validated with DMRs results from HNSCC TCGA samples. Significant DMRs common to both sample sets were subjected to Gene Ontology (GO) analysis with Database for Annotation, Visualization and Integrated Discovery version 6.7 (https://david.ncifcrf.gov/tools.jsp).

Candidate gene selection

Based on the genome wide DMR analysis, TCGA comparison and GO bioinformatics analysis, the 10 genes DAPK1, PITX2, PAX5, TIMP3, SFRP1, CALCA, SOCS1, CDH1, MAGI2 and PAX1 were selected for downstream validation as candidate biomarker genes for HNSCC, using quantitative methylation-specific PCR (qMSP).

DNA bisulfite modification

Bisulfite modification was used to convert unmethylated cytosine residues of genomic DNA into uracil while leaving the methylated cytosines unchanged. For all HNSCC and healthy samples, including discovery and prevalence cohort, 1 µg genomic DNA was treated with sodium bisulfite using the EZ DNA Methylation-Gold Kit (catalog no. D5005; Zymo Research Corp.) according to the manufacturer's protocol.

qMSP

All HNSCC and healthy DNA samples, including the discovery and prevalence cohorts, were subjected to qMSP. Tumor and healthy bisulfite-modified DNA samples were used as a qMSP assay template, a fluorescence-based real-time PCR assay was previously described (35). Primers and probe set sequences selected had been previously described to amplify the promoter regions of DAPK1, PITX2, PAX5, TIMP3, SFRP1, CALCA, SOCS1, CDH1, MAGI2 and PAX1, and a reference gene, ACTβ. Primers and probe set sequences are shown in Table SII (36–43). All qMSPs were carried out in duplicates in a 48-well reaction plate with a final volume of 25 µl. Each reaction contained 600 nM forward and reverse primers, 200 nM probe (Integrated DNA Technologies, Inc.), 1× TaqMan Universal PCR Master Mix, no UNG (Thermo Fisher Scientific, Inc.) and 2 µl bisulfite-modified DNA. qMSP amplifications were performed in a StepOne Real-Time PCR System (Thermo Fisher Scientific, Inc.) using the following conditions: 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and an annealing temperature of 58°C for 1 min. Each reaction plate included HNSCC bisulfite-modified DNA samples, a positive, fully methylated DNA control sample (bisulfite-converted Universal Methylated Human DNA Standard; Zymo Research Corp.) and no-template controls. Serial dilutions (30–0.003 ng) of bisulfite-converted Universal Methylated Human DNA standard were used to construct a calibration curve for each plate. After amplification, the percentage of methylated reference (PMR) for each candidate gene in each sample was calculated using the following equation: [(HNSCC sample Cq value gene of interest/HNSCC sample Cq value β-actin)/(fully methylated sample Cq value gene of interest/fully methylated sample Cq value β-actin] ×100.

PMR values obtained from samples from the discovery cohort were used to draw receiver operating characteristic (ROC) curves to obtain sensitivity and specificity values for every candidate gene. ROC curves were drawn using STATA (version 15; StataCorp LP). Based on sensitivity and specificity values, a suitable PMR cut-off value was chosen for every candidate gene. Prevalence cohort samples were classified as methylated (M) or unmethylated (UM) based on the PMR cut-off value for every candidate gene. Promoter methylation of PITX2, PAX5 and TIMP3 was tested in 29 OSCC samples. Promoter methylation of SFRP1, CALCA and SOCS1 was tested in 19 OPSCC samples. Promoter methylation of CDH1, MAGI2 and PAX1 was tested in 39 LSCC samples. Promoter methylation of DAPK1 was used as an internal control and was evaluated in all HNSCC samples.

Statistical analysis

Data from independent groups were compared using Fisher's exact test or χ²-test, as appropriate. Odds ratio (OR) calculations for clinicopathological parameters were performed using binary logistic. OS was measured in months from the date of diagnosis until death (if applicable). Survival analyses were performed using Kaplan-Meier curves. Log-rank Mantel-Cox and Gehan-Breslow Wilcoxon tests were used to determine the significance between two survival curves. Prognostic factors that have impact on HNSCC survival were analyzed in a Cox regression analysis. Statistical analyses were performed using SPSS (version 22; IBM Corp.). P<0.05 was considered to indicate a statistically significant difference.

Results <sec> <title>DMRs in HNSCC tumor samples from the discovery cohort

Results from the discovery cohort show that the three HNSCC subsites had in common 2,565 DMRs that included genes previously associated with HNSCC (Table SIII). Some of the identified DMRs corresponded to genes previously described as having a pivotal role in HNSCC carcinogenesis, such as BRCA2, CDKN2A, CDKN1B (P27), DAPK1,MAPK1, MAPK10, MLH1, RASSF1, HOXC6, VEGFB, WNT1 and WNT8B (44–55). Among these genes, several of them have roles in essential pathways for cell cycle regulation (RASSF1 and CDKN1B), cell proliferation (MAPK1 and MAPK10) and apoptosis (DAPK1).

The genome-wide analysis also unveiled 889 DMRs unique for OSCC, 363 DMRs for OPSCC and 738 DMRs for LSCC (Fig. S2). Results from the 450K Infinium DNA methylation array from 274 HNSCC TCGA samples and 32 frequency-matched UPPP control samples from John Hopkins Head and Neck Cancer Research Laboratory were used to validate subsite-specific DMRs identified in the MeDIP experiment. A GO analysis was used to describe the function of the most critical DMRs. Based on the DMR and GO analyses, DAPK1, PITX2, PAX5, TIMP3, SFRP1, CALCA, SOCS1, CDH1, MAGI2 and PAX1 were selected as candidate genes to be further evaluated.

The promoter methylation status of the 10 candidate genes in all HNSCC and healthy samples from the discovery cohort was analyzed. Samples were subjected to qMSP analysis for all candidate genes. A total of nine candidate genes showed differential methylation between HNSCC and healthy samples. The candidate gene SOCS1 showed no difference in the promoter methylation status between HNSCC and healthy samples. Table II shows values obtained for sensitivity, specificity, ROC curve and PMR cut-off value for every candidate gene.

In the OSCC samples, M PITX2 and PAX5 were detected in 58.6 and 79.3% of the samples, respectively. Also, M TIMP3 was confirmed in 79.3% of the samples, and M DAPK1 was detected in 51.7% of the samples. In the OPSCC samples, M SFRP1, CALCA and SOCS1 were detected in 84.2, 78.9 and 15.8% of the samples, respectively. M DAPK1 was detected in 63.2% of the samples. As for LSCC samples, M CDH1, MAGI1 and PAX1 were detected in 58.9, 64.1 and 74.4% of the samples, respectively. M DAPK1 was detected in 66.7% of the LSCC samples.

A total of five candidate genes were selected, DAPK1, CDH1, PAX1, CALCA and TIMP3, for validation in a HNSCC prevalence cohort based on predictive accuracy of the genes. Each candidate gene's predictive accuracy to detect HNSCC was calculated by ROC curve analysis. The sensitivity and specificity values were used to select the PMR cut-off value for every candidate gene. PAX1 and TIMP3 had ROC values of 1.00, 100% sensitivity and 100% specificity; DAPK1 had a ROC value of 0.92, 88.9% sensitivity and 100% specificity; CALCA had a ROC value of 0.90, 85.7% sensitivity and 100% specificity; and CDH1 had a ROC value of 0.82, 75% sensitivity and 90% specificity.

A PMR value was calculated for every candidate gene for all HNSCC and healthy samples in the prevalence cohort. PMR values obtained from the prevalence cohort were compared with the PMR cut-off value for e every candidate gene obtained from the discovery cohort (Fig. 2). Prevalence cohort samples with a ≥PMR value than the PMR cut-off value for every candidate gene were classified as M. Prevalence samples with a lower PMR than the PMR cut-off value for every candidate gene were classified as UM. The prevalence of M candidate genes in HNSCC and normal samples from the prevalence cohort is shown in Table SIV. Promoter aberrant methylation of DAPK1 was detected in 58.0% of the HNSCC samples (P=0.005), M CDH1 was detected in 50.0% of the HNSCC samples (P=0.112), methylation of PAX1 was confirmed in 82.0% of the HNSCC samples (P=0.001), M CALCA was confirmed in 44.0% of the HNSCC samples (P=0.036), while M TIMP3 was confirmed in 76.0% of the HNSCC prevalence cohort samples (P=0.003).

Association analysis between HNSCC clinicopathological characteristics and aberrant methylation of DAPK1, CDH1, PAX1, CALCA and TIMP3 (Table III) shows that the frequency of M PAX1 was significantly different across HNSCC anatomic subsites (P=0.029), being the highest frequency of detection in LSCC. No significant association was found between aberrant M genes (DAPK1, CDH1, CALCA and TIMP3) with sex, age, smoking, alcohol abuse, HPV infection and tumor staging. Concurrent methylation of two to five candidate genes was found in 15% of the patients with HNSCC (Fig. S3).

The prognostic value of DAPK1, CDH1, PAX1, CALCA and TIMP3 was assessed using Kaplan-Meier. Kaplan Meier survival analysis showed that patients with HNSCC and aberrant M DAPK1 had a better OS (61.0 months) compared with UM DAPK1 OS (24 months; P=0.043). No significant association with the OS of patients with HNSCC and aberrant methylation of CDH1, PAX1, CALCA and TIMP3 was found (Fig. 3).

A Cox regression analysis of OS was also performed to evaluate the association between aberrant methylation of the five candidate genes DAPK1, CDH1, PAX1, CALCA and TIMP3, and the risk of death from HNSCC (Table IV). Clinicopathological indicators such as age, tumor stage and differentiation, smoking, drinking and HPV infection were included in the analysis to assess their effect on HNSCC survival outcomes in this cohort. DAPK methylation (P=0.001; HR, 0.096; 95% CI, 0.03–0.37); HPV⁻ tumors (P=0.006; HR, 9.72; 95% CI, 1.94–48.42); tumor site (P=0.033; HR, 0.96; 95% CI, 0.03–0.37); tumor differentiation (P=0.013; HR, 8.56; 95% CI, 1.58–46.46); and age (P=0.002; HR, 1.11; 95% CI, 1.04–1.18) showed significant HRs.

Discussion

HNSCC is a heterogeneous disease comprising tumors from multiple anatomic subsites, each differing in prognosis and treatment strategy. Aberrant DNA methylation changes have been shown useful as molecular classifiers in several tumor sites because of their predictive capacity for disease detection, patient prognosis and treatment response (56–58). The discovery of epigenetic alterations is critical for a better understanding of HNSCC initiation and progression. Thus, the identification of genes epigenetically inactivated as potential prognostic biomarkers for HNSCC is urgent.

In the current study, an epigenomic analysis of a well-defined HNSCC cohort was shown. A genome-wide DNA methylation analysis showed that HNSCC tumors have 2,565 DMRs common to all HNSCC subsites. Several critical DMRs associated with a specific HNSCC anatomical subsite were identified. A total of 889, 363 and 738 DMRs unique to OSCC, OPSCC and LSCC were identified, respectively. These DMRs were associated with critical cellular pathways, often deregulated in multiple cancer types, including HNSCC (59). Among those DMRs, 10 candidate genes (DAPK1, PITX2, PAX5, TIMP3, SFRP1, CALCA, SOCS1, CDH1, MAGI2 and PAX1) were selected and evaluated further for their predictive and prognostic value.

Kaplan-Meier survival analysis (P=0.043) and Cox regression analysis of OS (P=0.001) showed that DAPK1 methylation is associated with better prognosis in HNSCC. DAPK1, a mediator of a wide range of cellular processes including growth, apoptosis, autophagy and oxidative stress (60,61), was identified as aberrantly methylated in all HNSCC subsites. DAPK1 prediction analysis suggested high levels of specificity and sensitivity for HNSCC detection, which was later confirmed in the HNSCC prevalence study (P=0.005). Epigenetic inactivation of DAPK1 may be a key event in head and neck carcinogenesis (62). Aberrant methylation of DAPK1 has been confirmed in the cancer of the OSCC (63–65), pharynx (66) and larynx (67). Aberrant methylation of DAPK1 has been shown to occur in advanced HNSCC (stages III and IV) tumors with positive lymph node involvement, resulting in a poor prognosis (67). Likewise, downregulation of DAPK1 expression has also been shown in HNSCCs (68,69). Loss of DAPK1 expression, mediated by promoter hypermethylation, has been associated with deregulation of autophagy in cancer cells and resistance to radiotherapy and chemotherapy treatments (70). Thus, aberrantly methylated DAPK1 may be associated with HNSCC carcinogenesis and progression.

Preliminary data from our research group suggested that aberrantly M CDH1 was associated with worse outcome (death) in LSCC. The present study shows that CDH1 is frequently M in LSCC. CDH1, a tumor invasion/suppressor gene, transcribes a 120-kDa glycoprotein, E-cadherin (71), that is essential for establishing and maintaining intercellular connections (72). Squamous carcinoma cells are characterized by poor cellular adhesion, loss of epithelial morphology and increased cellular motility (73). Downregulation of E-cadherin expression, either by genetic mutation or epigenetic dysregulation, leads to alterations in cell-to-cell adhesion and increases the metastatic potential of squamous cell carcinoma. M CDH1 has been associated with invasive LSCC tumors (Grade 3 and 4) and metastasis (74). M CDH1 has also been detected in tissue samples of surrounding mucosa of OSCC, suggesting that M CDH1 may be a key contributor to HNSCC carcinogenesis. Downregulation of CDH1 expression has been observed in HNSCC, and loss of CDH1 expression was associated with invasive HNSCC (75). Low CDH1 mRNA levels were detected in patients with tongue cancer (76), but no statistically significant association with clinicopathological characteristics nor patient outcome were found. A meta-analysis of 23 studies showed that CDH1 methylation was notably more frequent in HNSCC tissue than healthy controls, thus, supporting the role of CDH1 as a diagnostic biomarker (77). That meta-analysis showed that Asians display a higher frequency of MCDH1 than Caucasian or African subgroups and suggested that ethnicity may account for the differences in CDH1 methylation frequency (77). In the present study, it was confirmed that CDH1 is methylated at high levels in LSCC tumors and could act as a promising prognostic biomarker of LSCC.

The findings of the current study suggest that M PAX1 is also a promising biomarker for LSCC. M PAX1 was mostly detected in LSCC and was significantly different than control samples (P≤0.001). The frequency of M PAX1 was significantly different across HNSCC subsites, showing a higher frequency in LSCC (P=0.029) corroborating the genome-wide analysis. The results of the present study show that aberrant methylation of PAX1 had predictive accuracy for identifying HNSCC samples (ROC, 1.00; 100% specificity and sensitivity). PAX1 belongs to the highly conserved PAX gene family, which are developmentally controlled and encode transcription factors regulating embryogenesis in vertebrates (78). The expression of PAX1 during the development process is limited to the skeleton, thymus and parathyroid glands (79). PAX1, among other PAX genes, has critical roles in the development of the thymus and the parathyroid gland (80). PAX1 dysregulation causes a hypoplastic thymus with defects in thymocyte maturation and a delay in separation from the oropharynx (81), suggesting that loss of PAX1 dysregulates proliferation of the thymus (82). Loss of PAX1 expression, mediated by aberrant promoter methylation, has been detected in multiple cancer types, including cervical, colorectal and esophageal cancer, and in HNSCC (83,84). It has been shown that PAX1 is aberrantly M in HSCCC, including PAX1 aberrant methylation association with a higher risk of HNSCC (85,86). PAX1 methylation has been mostly studied in OSCC and associated with larger tumor size (87–89).

PAX1 methylation has been studied in HPV⁺ cervical cancer in which HPV infection disrupts epigenetic regulation through a series of aberrant methylation changes in the host genome (90). Likewise, HPV-induced aberrant methylation may affect the carcinogenic activity and clinical manifestation of HPV⁺ HNSCC and distinguish it from HPV⁻ HNSCC. Currently in Taiwan, M PAX1 is used for cervical cancer screening due to its association with increasing cervical dysplasia. Likewise, the association of M PAX1 with HPV infection in cervical cancer suggests that HPV may regulate PAX1 methylation; thus, further analysis of the significance of M PAX1 in HPV⁺ patients with HNSCC is warranted.

Among other candidate genes evaluated, CALCA, a potent vasodilator, and an essential inflammatory response molecule (91), was predicted as particularly M in OPSCC. Detection of aberrantly M CALCA in HNSCC showed to be highly accurate with 100% specificity and 85% sensitivity (P=0.0006). M CALCA was detected in 79% of the 19 OPSCC tumor samples, corroborating the predictive analysis. Additional assessment in 50 HNSCC samples showed that M CALCA was predominant in 48% of HNSCC samples (P=0.036). M CALCA has been detected in leukemia, testicular, bladder, non-small cell lung, thyroid, and head and neck cancer (92–97). In leukemia, M CALCA has been associated with disease relapse and poor prognosis (98,99). Likewise, studies in non-small lung carcinoma reveal that M CALCA was more common in squamous cell carcinomas than adenocarcinomas (100). Furthermore, M CALCA was associated with a poor prognosis in non-small lung carcinoma independent of the tumor stage (101). Aberrant methylation of CALCA in HNSCC has been studied mostly in oral cancer samples. Guerrero-Preston et al (102) showed a higher frequency of M CALCA in OSCC, although the prognostic value for HNSCC was not discussed. An additional study showed that M CALCA was associated with metastasis and a poorer prognosis in OSCC (103). Thus, the prognostic value of CALCA in HNSCC demands further study, particularly in OPSCC.

M TIMP3, a tissue inhibitor of metalloproteinase 3, was significantly detected in HNSCC (76%; P=0.003). The frequency of M TIMP3 was higher in OSCC (85%) compared with other subsites and corroborated the genome-wide analysis. Previous studies have shown M TIMP3 in HNSCC (104) in addition to various tumor types, including kidney, brain and esophageal cancer (105,106). TIMP3 is a critical regulator of inflammation (107,108), and loss of TIMP3 expression has been associated with an increase in cell proliferation, tumor growth, angiogenesis and metastasis (109–112). No significant association was identified between M TIMP3 and HNSCC clinical features in the cohort of the present study. Previous studies have detected M TIMP3 more frequently in tumor tissue of patients with early-stage (I/II) HNSCC compared with healthy saliva samples (113,114) but they did not find a clinical association with cancer. A study characterizing the epigenome in HPV⁺ patients with HNSCC showed that aberrant TIMP3 was predominant in HPV⁺ patients (115). The authors showed that M TIMP3 was associated with positive lymph node spread, and consequently, with a poorer prognosis. Therefore, further studies of M TIMP3 in HPV⁺ and HPV⁻ patients with HNSCC are warranted.

The current study also assessed the aberrant methylation of SFRP1, MAGI2, PITX2 and PAX5 in HNSCC. In OSCC, M PAX5 was detected in 79.3% of tumor samples, and M PITX2 was detected in 58.6% of OSCC samples. M MAGI2 was detected in 64.1% of LSCCs, and aberrant M SFRP1 was detected in 84.2% of oroOPSCC samples. The predictive accuracy of SFRP1, MAGI2, PITX2 and PAX5 was significant, and it should be further explored in a larger HNSCC cohort to assess their predictive value accurately.

Cox regression analysis of OS revealed that HPV⁻ patients with HNSCC were at a higher risk of dying of cancer (P=0.006; HR, 9.72; 95% CI, 1.94–48.42) compared to HPV⁺ patients with HNSCC. Previous studies have shown that HPV⁺ tumors have higher methylation levels due to an overexpression of DNMT3a (116–118). HPV⁺ tumors commonly inactivate CDKN2A by hypermethylation, whereas HPV⁻ tumors mostly inactivate CDKN2A by deletions or mutations (119). Likewise, studies have shown that HPV⁻ tumors have a global hypomethylation state and higher genomic instability compared with HPV⁺ tumors (120–122). It has been proposed that the machinery of HPV⁺ cells re-establish developmental methylation patterns as a defense mechanism to abolish transcription of the virus (122), which results in increased DNA methylation. Differences in DNA methylation patterns between HPV⁺ and HPV⁻ HNSCC must be substantiated and analyzed for their potential clinical translation into targeted treatment options.

In summary, the findings of the current study suggest that distinctive aberrant DNA methylation profiles arise within every head and neck cancer anatomic subsite, thus explaining the disease heterogeneity and possible association with disease progression or response to treatment. It was also shown that DAPK1, CDH1, PAX1, CALCA and TIMP3 are frequently aberrantly M in patients with head and neck cancer and influence survival. PAX1 hypermethylation was different across HNSCC anatomic subsites (P=0.029), and predominantly detected in LSCC. Kaplan-Meier survival analysis (P=0.043) and Cox regression analysis of OS (P=0.001) showed that DAPK1 methylation is associated with better prognosis in HNSCC.

The main limitation of the current study was the small sample size. Further studies with larger cohorts are needed to validate the results obtained. Determination of the global epigenetic landscape of HNSCC will require large cohorts of samples and coordinated research efforts to predict better biomarkers for disease outcome and targeted therapeutic interventions. Furthermore, epigenomic studies in which HNSCC tumors are stratified by anatomic site and HPV positivity are needed to better understand the modulation of the host epigenome by HPV modifying, and subsequent impact in disease progression and severity.

Supplementary Material

Supporting Data

Acknowledgements

Not applicable.

Availability of data and materials

The data that support the findings of this study are available from LifeGeneBiomarks, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of LifeGeneBiomarks.

Authors' contributions

BRP, RGP, DS and AB performed the study design. AB and JAAO obtained informed consent from patients with HNSCC and healthy individuals. BRP and AB collected all HNSCC tumor and healthy oral samples. BRP, RV, JAAO and RL carried out DNA extraction from HNSCC and healthy samples. BRP, OF and FP carried out the MeDIP assay, hybridization and scanning of samples to the microarray. BRP, NT and RGP performed bioinformatics analysis. RJRB and MEF analyzed the data. BRP, NT, SRT and RGP performed GO analysis, TCGA and Cancer Genome Browser analysis. DS financially supported this study and revised and approved final manuscript. BRP, DS and RGP confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

All procedures described in the present study were approved by the University of Puerto Rico-Medical Sciences Campus IRB (approval no. MSC-IRB Protocol 2770103). Written informed consent was obtained from all study participants.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Abbreviations

HNSCC

head and neck squamous cell carcinoma

HPV16

human papillomavirus 16

LSCC

larynx squamous cell carcinoma

OSCC

oral cavity squamous cell carcinoma

OPSCC

oropharynx squamous cell carcinoma

DMRs

differentially methylated regions

overall survival

qPCR

quantitative polymerase chain reaction

TSGs

tumor suppressor genes

MeDIP

methylated DNA immunoprecipitation

TSS

transcription starting site

TCGA

The Cancer Genome Atlas

Gene Ontology

IP DNA

immunoprecipitated DNA

WGA

whole genome amplification

References 1

Bray

Ferlay

Soerjomataram

Siegel

Torre

Jemal

Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries

CA Cancer J Clin683944242018

10.3322/caac.21492

30207593

O'Rorke

Ellison

Murray

Moran

James

Anderson

Human papillomavirus related head and neck cancer survival: A systematic review and meta-analysis

Oral Oncol48119112012012

10.1016/j.oraloncology.2012.06.019

22841677

Báez

Genetic and environmental factors in head and neck cancer genesis

J Environ Sci Health C Environ Carcinog Ecotoxicol Rev261742002008

10.1080/10590500802129431

18569329

Saracci

The interactions of tobacco smoking and other agents in cancer etiology

Epidemiol Rev91751931987

10.1093/oxfordjournals.epirev.a036301

3315716

Blot

McLaughlin

Winn

Austin

Greenberg

Preston-Martin

Bernstein

Schoenberg

Stemhagen

Fraumeni

Smoking and drinking in relation to oral and pharyngeal cancer

Cancer Res48328232871988

3365707

Hashibe

Brennan

Benhamou

Castellsague

Chen

Curado

Dal Maso

Daudt

Fabianova

Fernandez

Alcohol drinking in never users of tobacco, cigarette smoking in never drinkers, and the risk of head and neck cancer: Pooled analysis in the International Head and Neck Cancer Epidemiology Consortium

J Natl Cancer Inst997777892007

10.1093/jnci/djk179

17505073

Gillison

Koch

Capone

Spafford

Westra

Zahurak

Daniel

Viglione

Symer

Evidence for a causal association between human papillomavirus and a subset of head and neck cancers

J Natl Cancer Inst927097202000

10.1093/jnci/92.9.709

10793107

Mork

Lie

Glattre

Hallmans

Jellum

Koskela

Møller

Pukkala

Schiller

Youngman

Human papillomavirus infection as a risk factor for squamous-cell carcinoma of the head and neck

N Engl J Med344112511312001

10.1056/NEJM200104123441503

11297703

Báez

Almodóvar

Cantor

Celestin

Cruz-Cruz

Fonseca

Trinidad-Pinedo

Vega

High frequency of HPV16-associated head and neck squamous cell carcinoma in the Puerto Rican population

Head Neck267787842004

10.1002/hed.20046

15350023

Rivera-Peña

Ruíz-Fullana

Vélez-Reyes

Rodriguez-Benitez

Marcos-Martínez

Trinidad-Pinedo

Báez

HPV-16 infection modifies overall survival of Puerto Rican HNSCC patients

Infect Agent Cancer11472004

10.1186/s13027-016-0095-4

27559359

Siegel

Miller

Jemal

Cancer statistics, 2017

CA Cancer J Clin677302017

10.3322/caac.21395

28055103

Park

Alabaster

Shen

Mell

Katzel

Undertreatment of women with locoregionally advanced head and neck cancer

Cancer125303330392019

10.1002/cncr.32187

31090932

Katzel

Merchant

Chaturvedi

Silverberg

Contribution of demographic and behavioral factors on the changing incidence rates of oropharyngeal and oral cavity cancers in northern California

Cancer Epidemiol Biomarkers Prev249789842015

10.1158/1055-9965.EPI-14-1416

25956592

Dayyani

Etzel

Liu

Lippman

Tsao

Meta-analysis of the impact of human papillomavirus (HPV) on cancer risk and overall survival in head and neck squamous cell carcinomas (HNSCC)

Head Neck Oncol2152010

10.1186/1758-3284-2-15

20587061

Shiboski

Schmidt

Jordan

Tongue and tonsil carcinoma: Increasing trends in the U.S. population ages 20–44 years

Cancer103184318492005

10.1002/cncr.20998

15772957

American Joint Committee on CancerAJCC Cancer Staging Manual8

New York

Springer

2016

Baliga

Yildiz

Bazan

Palmer

Jhawar

Konieczkowski

Grecula

Blakaj

Mitchell

Henson

Disparities in survival outcomes among Racial/Ethnic minorities with head and neck squamous cell cancer in the united states

Cancers (Basel)1517812023

10.3390/cancers15061781

36980667

Suárez

Calo

Hernández

Diaz

Figueroa

Ortiz

Age-standardized incidence and mortality rates of oral and pharyngeal cancer in Puerto Rico and among Non-Hispanics Whites, Non-Hispanic Blacks, and Hispanics in the USA

BMC Cancer91292009

10.1186/1471-2407-9-129

19400958

Fernandez

Assenov

Martin-Subero

Balint

Siebert

Taniguchi

Yamamoto

Hidalgo

Tan

Galm

A DNA methylation fingerprint of 1628 human samples

Genome Res224074192012

10.1101/gr.119867.110

21613409

Taby

Issa

Cancer epigenetics

CA Cancer J Clin603763922010

10.3322/caac.20085

20959400

Ibáñez de Cáceres

Cairns

Methylated DNA sequences for early cancer detection, molecular classification and chemotherapy response prediction

Clin Transl Oncol94294372007

10.1007/s12094-007-0081-9

17652056

Calmon

Colombo

Carvalho

Souza

Filho

Fukuyama

Camargo

Caballero

Tajara

Cordeiro

Rahal

Methylation profile of genes CDKN2A (p14 and p16), DAPK1, CDH1, and ADAM23 in head and neck cancer

Cancer Genet Cytogenet17331372007

10.1016/j.cancergencyto.2006.09.008

17284367

Demokan

Dalay

Role of DNA methylation in head and neck cancer

Clin Epigenetics21231502011

10.1007/s13148-011-0045-3

22704334

Ovchinnikov

Cooper

Pandit

Coman

Cooper-White

Keith

Wolvetang

Slowey

Punyadeera

Tumor-suppressor gene promoter hypermethylation in saliva of head and neck cancer patients

Transl Oncol53213262012

10.1593/tlo.12232

23066440

Herman

Baylin

Gene silencing in cancer in association with promoter hypermethylation

N Engl J Med349204220542003

10.1056/NEJMra023075

14627790

Esteller

Epigenetics in cancer

N Engl J Med358114811592008

10.1056/NEJMra072067

18337604

Lleras

Smith

Adrien

Schlecht

Burk

Harris

Childs

Prystowsky

Belbin

Unique DNA methylation loci distinguish anatomic site and HPV status in head and neck squamous cell carcinoma

Clin Cancer Res19544454552013

10.1158/1078-0432.CCR-12-3280

23894057

Bernabe

INK4a/ARF/INK4b tumor suppressor locus: Its potential role in head and neck cancer tumorigenesis (Doctoral dissertation) (order no. 3221909)

ProQuest Dissertations and Theses Global2006Available fromhttps://search.proquest.com/docview/304984294?accountid=44820

Rivera-Peña

Báez

Abstract 4793: Aberrant methylation of CDH1 correlates with poor prognosis in patients with head and neck squamous cell carcinoma (abstract). In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research

Cancer Res71(Suppl 8)479347932011

10.1158/1538-7445.AM2011-4793

Irizarry

Ladd-Acosta

Carvalho

Brandenburg

Jeddeloh

Wen

Feinberg

Comprehensive high-throughput arrays for relative methylation (CHARM)

Genome Res187807902008

10.1101/gr.7301508

18316654

Jaffe

Murakami

Lee

Leek

Fallin

Feinberg

Irizarry

Bump hunting to identify differentially methylated regions in epigenetic epidemiology studies

Int J Epidemiol412002092012

10.1093/ije/dyr238

22422453

Zhang

Huang

Jafari

Kibbe

Hou

Lin

Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis

BMC Bioinformatics115872010

10.1186/1471-2105-11-587

21118553

The Cancer Genome Atlas Network

Comprehensive genomic characterization of head and neck squamous cell carcinomas

Nature5175765822015

10.1038/nature14129

25631445

Aryee

Jaffe

Corrada-Bravo

Ladd-Acosta

Feinberg

Hansen

Irizarry

Minfi: A flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays

Bioinformatics30136313692014

10.1093/bioinformatics/btu049

24478339

Eads

Danenberg

Kawakami

Saltz

Blake

Shibata

Danenberg

Laird

MethyLight: A high-throughput assay to measure DNA methylation

Nucleic Acids Res28E322000

10.1093/nar/28.8.e32

10734209

Rettori

de Carvalho

Longo

de Oliveira

Kowalski

Carvalho

Vettore

TIMP3 and CCNA1 hypermethylation in HNSCC is associated with an increased incidence of second primary tumors

J Transl Med113162013

10.1186/1479-5876-11-316

24359512

Dahiya

MethPrimer: Designing primers for methylation PCRs

Bioinformatics18142714312002

10.1093/bioinformatics/18.11.1427

12424112

Valle

Rodriguez-Torres

Kuhn

Díaz-Montes

Parrilla-Castellar

Lawson

Folawiyo

Ili-Gangas

Brebi-Mieville

Eshleman

HIST1H2BB and MAGI2 methylation and somatic mutations as precision medicine biomarkers for diagnosis and prognosis of high-grade serous ovarian cancer

Cancer Prev Res (Phila)137837942020

10.1158/1940-6207.CAPR-19-0412

32581010

Guerrero-Preston

Valle

Jedlicka

Turaga

Folawiyo

Pirini

Lawson

Vergura

Noordhuis

Dziedzic

Molecular triage of premalignant lesions in Liquid-Based cervical cytology and circulating Cell-Free DNA from urine, using a panel of methylated human papilloma virus and host genes

Cancer Prev Res (Phila)99159242016

10.1158/1940-6207.CAPR-16-0138

27671338

Guerrero-Preston

Michailidi

Marchionni

Pickering

Frederick

Myers

Yegnasubramanian

Hadar

Noordhuis

Zizkova

Key tumor suppressor genes inactivated by ‘greater promoter’ methylation and somatic mutations in head and neck cancer

Epigenetics9103110462014

10.4161/epi.29025

24786473

Anglim

Galler

Koss

Hagen

Turla

Campan

Weisenberger

Laird

Siegmund

Laird-Offringa

Identification of a panel of sensitive and specific DNA methylation markers for squamous cell lung cancer

Mol Cancer7622008

10.1186/1476-4598-7-62

18616821

Müller

Widschwendter

Fiegl

Ivarsson

Goebel

Perkmann

Marth

Widschwendter

DNA methylation in serum of breast cancer patients: An independent prognostic marker

Cancer Res63764176652003

14633683

Eads

Lord

Wickramasinghe

Long

Kurumboor

Bernstein

Peters

DeMeester

Skinner

Laird

Epigenetic patterns in the progression of esophageal adenocarcinoma

Cancer Res61341034182001

11309301

van Asperen

Brohet

Meijers-Heijboer

Hoogerbrugge

Verhoef

Vasen

Ausems

Menko

Gomez Garcia

Klijn

Cancer risks in BRCA2 families: Estimates for sites other than breast and ovary

J Med Genet427117192005

10.1136/jmg.2004.028829

16141007

Seiwert

Zuo

Keck

Khattri

Pedamallu

Stricker

Brown

Pugh

Stojanov

Cho

Integrative and comparative genomic analysis of HPV-positive and HPV-negative head and neck squamous cell carcinomas

Clin Cancer Res216326412015

10.1158/1078-0432.CCR-13-3310

25056374

Califano

Van Der Riet

Westra

Nawroz

Clayman

Piantadosi

Corio

Lee

Greenberg

Koch

Sidransky

Genetic progression model for head and neck cancer: Implications for field cancerization

Cancer Res56248824921996

8653682

Cai

Xiao

Niu

Zhong

Association between promoter methylation of DAPK gene and HNSCC: A meta-analysis

PLoS One12e01731942017

10.1371/journal.pone.0173194

28249042

Ngan

Liu

Fong

Poon

PHY

Yeung

Chan

SSM

Lau

Piao

Tse

MAPK pathway mutations in head and neck cancer affect immune microenvironments and ErbB3 signaling

Life Sci Alliance3e2019005452020

10.26508/lsa.201900545

32381551

Cilona

Locatello

Novelli

Gallo

The mismatch repair system (MMR) in head and neck carcinogenesis and its role in modulating the response to immunotherapy: A critical review

Cancers (Basel)1230062020

10.3390/cancers12103006

33081243

Meng

Chen

Huang

Shi

Niu

Aberrant Methylation of RASSF1A closely associated with HNSCC, a Meta-Analysis

Sci Rep6207562016

10.1038/srep20756

26857374

Moon

Kim

Yoon

Ahn

HOXC6 is deregulated in human head and neck squamous cell carcinoma and modulates Bcl-2 expression

J Biol Chem28735678356882012

10.1074/jbc.M112.361675

22896703

Carla

Daris

Cecilia

Francesca

Paolo

Angiogenesis in head and neck cancer: A review of the literature

J Oncol20123584722012

10.1155/2012/358472

22131994

Mineta

Miura

Ogino

Takebayashi

Misawa

Ueda

Suzuki

Dictor

Borg

Wennerberg

Prognostic value of vascular endothelial growth factor (VEGF) in head and neck squamous cell carcinomas

Br J Cancer837757812000

10.1054/bjoc.2000.1357

10952783

Iwai

Katagiri

Kong

Amekawa

Nakazawa

Yura

Mutations of the APC, beta-catenin, and axin 1 genes and cytoplasmic accumulation of beta-catenin in oral squamous cell carcinoma

J Cancer Res Clin Oncol1317737822005

10.1007/s00432-005-0027-y

16163548

Leethanakul

Patel

Gillespie

Pallente

Ensley

Koontongkaew

Liotta

Emmert-Buck

Gutkind

Distinct pattern of expression of differentiation and growthrelated genes in squamous cell carcinomas of the head and neck revealed by the use of laser capture microdissection and cDNA arrays

Oncogene19322032242000

10.1038/sj.onc.1203703

10918578

Paska

Hudler

Aberrant methylation patterns in cancer: A clinical view

Biochem Med (Zagreb)251611762015

10.11613/BM.2015.017

26110029

Hao

Luo

Krawczyk

Wei

Wang

Flagg

Hou

Zhang

DNA methylation markers for diagnosis and prognosis of common cancers

Proc Natl Acad Sci USA114741474192017

10.1073/pnas.1703577114

28652331

Chen

Wang

Guo

Dou

Tang

Zong

Zhang

Xiong

An aberrant DNA methylation signature for predicting the prognosis of head and neck squamous cell carcinoma

Cancer Med10593659472021

10.1002/cam4.3965

34313009

Stadler

Patel

Couch

Hayes

Molecular biology of head and neck cancer: Risks and pathways

Hematol Oncol Clin North Am22109911242008

10.1016/j.hoc.2008.08.007

19010262

Bialik

Kimchi

The death-associated protein kinases: Structure, function, and beyond

Annu Rev Biochem751892102006

10.1146/annurev.biochem.75.103004.142615

16756490

Gade

Manjegowda

Nallar

Maachani

Cross

Kalvakolanu

Regulation of the death-associated protein kinase 1 expression and autophagy via ATF6 requires apoptosis signal-regulating kinase 1

Mol Cell Biol3440334482014

10.1128/MCB.00397-14

25135476

Wang

Zhang

Zhou

mRNA expression and hypermethylation of tumor suppressor genes apoptosis protease activating factor-1 and death-associated protein kinase in oral squamous cell carcinoma

Oncol Lett62802862013

10.3892/ol.2013.1353

23946818

Jayaprakash

Varghese

Bellampalli

Radhakrishnan

Ray

Kabekkodu

Satyamoorthy

Hypermethylation of Death-associated protein kinase (DAPK1) and its association with oral carcinogenesis-An experimental and meta-analysis study

Arch Oral Biol801171292017

10.1016/j.archoralbio.2017.03.024

28412611

Melchers

Clausen

Mastik

Slagter-Menkema

van der Wal

Wisman

Roodenburg

Schuuring

Identification of methylation markers for the prediction of nodal metastasis in oral and oropharyngeal squamous cell carcinoma

Epigenetics108508602015

10.1080/15592294.2015.1075689

26213212

Wei

Liu

Lei

Jin

Wang

Pan

Aberrant methylation and expression of DAPk1 in human hypopharyngeal squamous cell carcinoma

Acta Otolaryngol13570782015

10.3109/00016489.2014.956335

25496179

van Kempen

van Bockel

Braunius

Moelans

van Olst

de Jong

Stegeman

van Diest

Grolman

Willems

HPV-positive oropharyngeal squamous cell carcinoma is associated with TIMP3 and CADM1 promoter hypermethylation

Cancer Med3118511962014

10.1002/cam4.313

25065733

Stephen

Chen

Shah

Havard

Kapke

Benninger

Worsham

DNA hypermethylation markers of poor outcome in laryngeal cancer

Clin Epigenetics161692010

10.1007/s13148-010-0005-3

21318053

Strzelczyk

Krakowczyk

Owczarek

Aberrant DNA methylation of the p16, APC, MGMT, TIMP3 and CDH1 gene promoters in tumours and the surgical margins of patients with oral cavity cancer

J Cancer9189619042018

10.7150/jca.24477

29896273

Schmezer

Plass

Epigenetic aspects in carcinomas of the head and neck

HNO565946022008(In German)

10.1007/s00106-008-1720-3

18483718

Singh

Ravanan

Talwar

Death associated protein kinase 1 (DAPK1): A regulator of apoptosis and autophagy

Front Mol Neurosci9462016

10.3389/fnmol.2016.00046

27445685

Humphries

Newham

The structure of cell-adhesion molecules

Trends Cell Biol878831998

10.1016/S0962-8924(98)80016-0

9695813

Riethmacher

Brinkmann

Birchmeier

A targeted mutation in the mouse E-cadherin gene results in defective preimplantation development

Proc Natl Acad Sci USA928558591995

10.1073/pnas.92.3.855

7846066

Pećina-Slaus

Tumor suppressor gene E-cadherin and its role in normal and malignant cells

Cancer Cell Int3172003

10.1186/1475-2867-3-17

14613514

Starska

Forma

Lewy-Trenda

Papież

Woś

Bryś

Diagnostic impact of promoter methylation and E-cadherin gene and protein expression levels in laryngeal carcinoma

Contemp Oncol (Pozn)172632712013

24596512

Fan

Wang

Cheng

Jiang

Cheng

Lee

Kao

Expression of E-cadherin, Twist, and p53 and their prognostic value in patients with oral squamous cell carcinoma

J Cancer Res Clin Oncol139173517442013

10.1007/s00432-013-1499-9

23990016

Fujii

Imanishi

Shibata

Sakai

Sakamoto

Shigetomi

Habu

Otsuka

Sato

Watanabe

Restoration of E-cadherin expression by selective Cox-2 inhibition and the clinical relevance of the epithelial-to-mesenchymal transition in head and neck squamous cell carcinoma

J Exp Clin Cancer Res33402014

10.1186/1756-9966-33-40

24887090

Shen

Zhou

Deng

Wang

The association, clinicopathological significance, and diagnostic value of CDH1 promoter methylation in head and neck squamous cell carcinoma: A meta-analysis of 23 studies

Onco Targets Ther9676367732016

10.2147/OTT.S117453

27826202

McGaughran

Oates

Donnai

Read

Tassabehji

Mutations in PAX1 may be associated with Klippel-Feil syndrome

Eur J Hum Genet114684742003

10.1038/sj.ejhg.5200987

12774041

Wallin

Eibel

Neubüser

Wilting

Koseki

Balling

Pax1 is expressed during development of the thymus epithelium and is required for normal T-cell maturation

Development12223301996

10.1242/dev.122.1.23

8565834

Yang

Wang

Lin

Assessing methylation status of PAX1 in cervical scrapings, as a novel diagnostic and predictive biomarker, was closely related to screen cervical cancer

Int J Clin Exp Pathol8167416812015

25973053

Dietrich

Gruss

Undulated phenotypes suggest a role of Pax-1 for the development of vertebral and extravertebral structures

Dev Biol1675295481995

10.1006/dbio.1995.1047

7875377

Ellis

Napier

Lee

Manley

Hoxa3 and pax1 regulate epithelial cell death and proliferation during thymus and parathyroid organogenesis

Dev Biol2363163292001

10.1006/dbio.2001.0342

11476574

Lorincz

Virtues and Weaknesses of DNA methylation as a test for cervical cancer prevention

Acta Cytol605015122016

10.1159/000450595

27806357

Juodzbalys

Kasradze

Cicciù

Sudeikis

Banys

Galindo-Moreno

Guobis

Modern molecular biomarkers of head and neck cancer. Part I. Epigenetic diagnostics and prognostics: Systematic review

Cancer Biomark174875022016

10.3233/CBM-160666

27802200

Huang

Peng

Wang

Lai

DNA methylation of PAX1 as a biomarker for oral squamous cell carcinoma

Clin Oral Investig188018082014

10.1007/s00784-013-1048-6

23907469

Sun

Juan

Zhang

Yang

Peng

Hypermethylated PAX1 and ZNF582 genes in the tissue sample are associated with aggressive progression of oral squamous cell carcinoma

J Oral Pathol Med497517602020

10.1111/jop.13035

32428271

Cheng

Chang

Lee

Chen

Wang

Lin

Chiang

Hypermethylated ZNF582 and PAX1 genes in mouth rinse samples as biomarkers for oral dysplasia and oral cancer detection

Head Neck403553682018

10.1002/hed.24958

28960639

Morandi

Gissi

Tarsitano

Asioli

Gabusi

Marchetti

Montebugnoli

Foschini

CpG location and methylation level are crucial factors for the early detection of oral squamous cell carcinoma in brushing samples using bisulfite sequencing of a 13-gene panel

Clin Epigenetics9852017

10.1186/s13148-017-0386-7

28814981

Cheng

Chang

Liu

Peng

Wang

Lin

Chiang

Hypermethylated ZNF582 and PAX1 genes in oral scrapings collected from cancer-adjacent normal oral mucosal sites are associated with aggressive progression and poor prognosis of oral cancer

Oral Oncol751691772017

10.1016/j.oraloncology.2017.11.013

29224816

Hsu

Huang

Chen

Wang

Liao

Lai

Genotype-specific methylation of HPV in cervical intraepithelial neoplasia

J Gynecol Oncol28e562017

10.3802/jgo.2017.28.e56

28541643

Brain

Williams

Tippins

Morris

MacIntyre

Calcitonin gene-related peptide is a potent vasodilator

Nature31354561985

10.1038/313054a0

3917554

Vidal

Paixão

Brait

Souto

Caballero

Lopes

Vettore

Aberrant methylation in pediatric myelodysplastic syndrome

Leuk Res311751812007

10.1016/j.leukres.2006.06.013

16890288

Martinelli

CMDS

Lengert

AVH

Cárcano

Silva

ECA

Brait

Lopes

Vidal

MGMT and CALCA promoter methylation are associated with poor prognosis in testicular germ cell tumor patients

Oncotarget850608506172016

10.18632/oncotarget.11167

28881587

Brait

Begum

Carvalho

Dasgupta

Vettore

Czerniak

Caballero

Westra

Sidransky

Hoque

Aberrant promoter methylation of multiple genes during pathogenesis of bladder cancer

Cancer Epidemiol Biomarkers Prev17278627942008

10.1158/1055-9965.EPI-08-0192

18843024

Wang

Zhang

Zheng

Luo

Bai

Multiple gene methylation of nonsmall cell lung cancers evaluated with 3-dimensional microarray

Cancer112132513362008

10.1002/cncr.23312

18286531

Zhang

Liu

Zhang

Wei

Yang

Hou

Shi

Analysis of BRAF(V600E) mutation and DNA methylation improves the diagnostics of thyroid fine needle aspiration biopsies

Diagn Pathol9452014

10.1186/1746-1596-9-45

24588959

Loyo

Brait

Kim

Ostrow

Jie

Chuang

Califano

Liégeois

Begum

Westra

Asurvey of methylated candidate tumor suppressor genes in nasopharyngeal carcinoma

Int J Cancer128139314032011

10.1002/ijc.25443

20473931

Ismail

El-Mogy

Mohamed

El-Farrash

Methylation pattern of calcitonin (CALCA) gene in pediatric acute leukemia

J Pediatr Hematol Oncol335345422011

10.1097/MPH.0b013e3181f46bc4

21423046

Paixão

Vidal

Caballero

Vettore

Tone

Ribeiro

Lopes

Hypermethylation of CpG island in the promoter region of CALCA in acute lymphoblastic leukemia with central nervous system (CNS) infiltration correlates with poorer prognosis

Leuk Res308918942006

10.1016/j.leukres.2005.11.016

16712930

100

Guan

Gao

Shi

Hou

Highly frequent promoter methylation and PIK3CA amplification in non-small cell lung cancer (NSCLC)

BMC Cancer111472011

10.1186/1471-2407-11-147

21507233

101

Morán

Fernández-Marcelo

Carro

De Juan

Pascua

Head

Gómez

Hernando

Torres

Benito

Iniesta

Methylation profiling in non-small cell lung cancer: Clinical implications

Int J Oncol407397462012

22076605

102

Guerrero-Preston

Soudry

Acero

Orera

Moreno-López

Macía-Colón

Jaffe

Berdasco

Ili-Gangas

Brebi-Mieville

NID2 and HOXA9 promoter hypermethylation as biomarkers for prevention and early detection in oral cavity squamous cell carcinoma tissues and saliva

Cancer Prev Res (Phila)4106110722011

10.1158/1940-6207.CAPR-11-0006

21558411

103

Jithesh

Risk

Schache

Dhanda

Lane

Liloglou

Shaw

The epigenetic landscape of oral squamous cell carcinoma

Br J Cancer1083703792013

10.1038/bjc.2012.568

23287992

104

Esteller

Corn

Baylin

Herman

A gene hypermethylation profile of human cancer

Cancer Res61322532292001

11309270

105

Bachman

Herman

Corn

Merlo

Costello

Cavenee

Baylin

Graff

Methylation-associated silencing of the tissue inhibitor of metalloproteinase-3 gene suggest a suppressor role in kidney, brain, and other human cancers

Cancer Res597988021999

10029065

106

Darnton

Hardie

Muc

Wild

Casson

Tissue inhibitor of metalloproteinase-3 (TIMP-3) gene is methylated in the development of esophageal adenocarcinoma: Loss of expression correlates with poor prognosis

Int J Cancer1153513582005

10.1002/ijc.20830

15688381

107

Smookler

Mohammed

Kassiri

Duncan

Mak

Khokha

Tissue inhibitor of metalloproteinase 3 regulates TNF-dependent systemic inflammation

J Immunol1767217252006

10.4049/jimmunol.176.2.721

16393953

108

Mohammed

Smookler

Taylor

Fingleton

Kassiri

Sanchez

English

Matrisian

Yeh

Khokha

Abnormal TNF activity in Timp3-/-mice leads to chronic hepatic inflammation and failure of liver regeneration

Nat Genet369699772004

10.1038/ng1413

15322543

109

Zhang

Zhao

Lin

Guo

Liang

Zhao

Fang

Inhibition of tumor growth and induction of apoptosis in prostate cancer cell lines by overexpression of tissue inhibitor of matrix metalloproteinase-3

Cancer Gene Ther171711792010

10.1038/cgt.2009.59

19798124

110

Ahonen

Ala-Aho

Baker

George

Grénman

Saarialho-Kere

Kähäri

Antitumor activity and bystander effect of adenovirally delivered tissue inhibitor of metalloproteinases-3

Mol Ther57057152002

10.1006/mthe.2002.0606

12027554

111

Spurbeck

Strom

Vanin

Davidoff

Enforced expression of tissue inhibitor of matrix metalloproteinase-3 affects functional capillary morphogenesis and inhibits tumor growth in a murine tumor model

Blood100336133682002

10.1182/blood.V100.9.3361

12384438

112

Bian

Wang

Smith

Kim

Jacobs

Jackman

Kung

Colburn

Sun

Suppression of in vivo tumor growth and induction of suspension cell death by tissue inhibitor of metalloproteinases (TIMP)-3

Carcinogenesis17180518111996

10.1093/carcin/17.9.1805

8824499

113

Arantes

de Carvalho

Melendez

Centrone

Góis-Filho

Toporcov

Caly

Tajara

Goloni-Bertollo

Carvalho AL;

GENCAPO

Validation of methylation markers for diagnosis of oral cavity cancer

Eur J Cancer516326412015

10.1016/j.ejca.2015.01.060

25686481

114

Righini

de Fraipont

Timsit

Faure

Brambilla

Reyt

Favrot

Tumor-specific methylation in saliva: A promising biomarker for early detection of head and neck cancer recurrence

Clin Cancer Res13117911852007

10.1158/1078-0432.CCR-06-2027

17317827

115

Zhang

Shan

Huang

Wang

Characterization of gene methylation in human papillomavirus associated-head and neck squamous cell carcinoma

Genet Mol Res152016doi: 10.4238/gmr.15038206

116

Sartor

Dolinoy

Jones

Colacino

Prince

Carey

Rozek

Genome-wide methylation and expression differences in HPV(+) and HPV(−) squamous cell carcinoma cell lines are consistent with divergent mechanisms of carcinogenesis

Epigenetics67777872011

10.4161/epi.6.6.16216

21613826

117

Nakagawa

Matsusaka

Misawa

Ota

Takane

Fukuyo

Rahmutulla

Shinohara

Kunii

Sakurai

Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer

Cancer Lett40721312017

10.1016/j.canlet.2017.08.008

28823962

118

Nakagawa

Kurokawa

Mima

Imamoto

Mizokami

Kondo

Okamoto

Misawa

Hanazawa

Kaneda

DNA Methylation and HPV-Associated Head and Neck Cancer

Microorganisms98012021

10.3390/microorganisms9040801

33920277

119

Nakahara

Shintani

Mihara

Ueyama

Matsumura

High frequency of homozygous deletion and methylation of p16(INK4A) gene in oral squamous cell carcinomas

Cancer Lett1632212282001

10.1016/S0304-3835(00)00699-6

11165758

120

Richards

Zhang

Baggerly

Colella

Lang

Schuller

Krahe

Genome-wide hypomethylation in head and neck cancer is more pronounced in HPV-negative tumors and is associated with genomic instability

PLoS One4e49412009

10.1371/journal.pone.0004941

19293934

121

van Kempen

Noorlag

Braunius

Stegeman

Willems

Grolman

Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: A systematic review

Epigenetics91942032014

10.4161/epi.26881

24169583

122

Camuzi

Buexm

Lourenço

SQC

Esposti

Cuenin

Lopes

MdSA

Manara

Talukdar

Herceg

Ribeiro Pinto

HPV infection leaves a DNA methylation signature in oropharyngeal cancer affecting both coding genes and transposable elements

Cancers1336212021

10.3390/cancers13143621

34298834

Figure 1.

Flow chart showing the integrated experimental design of the present study. HNSCC, head and neck squamous cell carcinoma; HPV, human papillomavirus; MeDIP, methylated DNA immunoprecipitation; CHARM, comprehensive high-throughput arrays for relative methylation; qMSP, quantitative methylation-specific PCR; PMR, percentage of methylated reference.

Figure 2.

Scatterplots of the PMR values obtained by qMSP in the analysis of selected candidate genes in the discovery cohort. The PMR for every candidate gene was expressed as a ratio of the amplification from the target gene to the amount amplified from the reference gene β-actin, then multiplied by 100. The black line represents the PMR cut-off value for each candidate gene. PMR, percentage of methylated reference; qMSP, quantitative methylation-specific PCR; LSCC, larynx squamous cell carcinoma; OSCC, oral cavity squamous cell carcinoma; OPSCC, oropharynx squamous cell carcinoma.

Figure 3.

Overall survival of patients with HNSCC according to the aberrant promoter methylation status of DAPK1, CDH1, PAX1, CALCA, and TIMP3. Kaplan-Meier survival curve analysis of patients with HNSCC and aberrantly meth genes vs. patients with HNSCC and unmeth genes. The black line represents patients with meth HNSCC and the dashed black line represents patients with unmeth HNSCC. Median survival in each curve is represented in months. Meth, methylated; unmeth, unmethylated; HNSCC, head and neck squamous cell carcinoma.

Table I.

Clinicopathological characteristics of the head and heck squamous cell carcinoma discovery cohort (n=21) and the prevalence cohort (n=119).

Characteristics	Discovery cohort	Prevalence cohort
Age, years	56.62±12.62	61.21±12.63
	(24–76)	(24–98)
Sex, n (%)
Male	19 (90.5)	107 (89.9)
Female	2 (9.5)	12 (10.1)
Site of primary tumor, n (%)
Oral cavity	10 (47.6)	42 (35.3)
Pharynx	4 (19.1)	25 (21.0)
Larynx	7 (33.3)	52 (43.7)
Tumor stage, n (%)^a
Early (I/II)	7 (33.3)	25 (21.0)
Advanced (III/IV)	14 (66.7)	92 (77.3)
HPV-16 status, n (%)^a
HPV-16⁺	7 (33.3)	57 (47.9)
HPV-16⁻	12 (57.1)	62 (52.1)
Differentiation, n (%)^b
Poor	2 (9.5)	9 (7.6)
Moderate	15 (71.4)	78 (65.5)
Well	4 (19.1)	31 (26.1)
Heavy smoking, n (%)	20 (95.2)	104 (87.4)
Heavy drinking, n (%)	21 (100.0)	97 (81.5)

Data was not available for 2 patients.

Data was not available for 1 patient. Data are shown as mean ± SD (range). HPV, human papillomavirus.

Table II.

Predictive accuracy of DAPK1, PITX2, PAX5, TIMP3, SFRP1, CALCA, SOCS1, CDH1, MAGI2 and PAX1 for head and neck squamous cell carcinoma.

Target genes	ROC	P-value	Sensitivity, %	Specificity, %	Methylation cut-off value
DAPK1	0.92	0.0009	88.89	100.00	12.36
PITX2	1.00	<0.0001	100.00	100.00	16.37
PAX5	0.96	<0.0001	88.89	100.00	15.71
TIMP3	1.00	<0.0001	100.00	100.00	4.52
SFRP1	0.96	0.0005	100.00	90.00	12.47
CALCA	0.90	0.0006	85.71	100.00	26.37
SOCS1	0.52	0.4750	57.14	90.00	4.06
CDH1	0.82	0.0054	75.00	90.00	16.15
MAGI2	0.96	0.0001	87.50	100.00	20.86
PAX1	1.00	<0.0001	100.00	100.00	29.41

ROC, receiver operating characteristic curve.

Table III.

Association between clinicopathological characteristics of the HNSCC cohort (n=50) and aberrant promoter methylation of DAPK1, CDH1, PAX1, CALCA and TIMP3.

	DAPK1					CDH1					PAX1					CALCA					TIMP3

Clinicopathological characteristics	Meth	Unmeth	Unknown	P-value	RR [95% CI]	Meth	Unmeth	Unknown	P-value	RR [95% CI]	Meth	Unmeth	Unknown	P-value	RR [95% CI]	Meth	Unmeth	Unknown	P-value	RR [95% CI]	Meth	Unmeth	Unknown	P-value	RR [95% CI]
Age, years
<60	10	7			1.02	9	8			1.09	15	2			1.12	6	11			0.72	11	6			0.79
≥60	19	14		1.000	[0.62-	16	17		1.000	[0.62-	26	7		0.699	[0.87-	16	17		0.548	[0.35-	27	6		0.294	[0.54-
					1.67]					1.93]					1.44]					1.51]					1.16]
Sex, n
Male	27	18			1.50	23	22			1.28	38	7			1.41	19	26			0.70	33	12			0.73
Female	2	3		0.638	[0.50-	2	3		1.000	[0.42-	3	2		0.216	[0.68-	3	2		0.642	[0.32-	5	0		0.319	[0.61-
					4.50]					3.88]					2.91]					1.55]					0.87]
HPV-16 status, n
HPV-16⁺	11	13			0.66	12	12			1.00	19	5			0.93	10	14			0.90	19	5			1.08
HPV-16⁻	18	8		0.151	[0.39-	13	13		1.000	[0.57-	22	4		0.721	[0.72-	12	14		0.783	[0.48-	19	7		0.745	[0.79-
					1.10]					1.74]					1.22]					1.69]					1.48]
Tumor site, n
Oral cavity	8	5				6	7				13	0				4	9				11	2
Pharynx	3	3				1	5				3	3				4	2				6	0
Larynx	18	13		0.894		18	13		0.169		25	6		0.029		14	17		0.334		21	10		0.167
Disease staging, n
Early (I/II)	6	5			0.96	6	5			1.12	11	0			1.32	4	7			0.79	10	1			1.29
Late (III/IV)	21	16	2	1.000	[0.52-	18	19	2	1.000	[0.59-	28	9	2	0.095	[1.10-	17	20	2	0.733	[0.34-	26	11	2	0.248	[0.98-
					1.77]					2.11]					1.59]					1.86]					1.71]
Nodal
involvement, n
Yes	7	3			1.29	5	5			1.02	7	3			0.84	4	6			0.87	6	4			0.77
No	20	17	3	0.481	[0.78-	18	19	3	1.000	[0.51-	31	6	3	0.377	[0.54-	17	20	3	1.000	[0.38-	29	8	3	0.251	[0.45-
					2.14]					2.08]					1.28]					2.01]					1.31]
Tumor
differentiation, n
Well	6	7				4	9				11	2				3	10				11	2
Moderate	19	10				17	13				25	5				14	16				22	8
Poor	2	4	2	0.243		3	3	1	0.296		4	2	1	0.597		4	2	1	0.162		4	2	1	0.633
Heavy smoking, n
Yes	25	18			1.74	23	20			8.02	35	8			1.22	19	24			0.66	32	10			1.14
No	1	2	4	0.572	[0.34-	0	3	4	0.233	[0.39-	2	1	4	0.488	[0.54-	2	1	4	0.585	[0.28-	2	1	5	1.000	[0.50-
					8.82]					164.9]					2.75]					1.58]					2.59]
Heavy drinking, n
Yes	22	16			1.16	21	17			2.21	32	6			1.35	16	22			0.67	28	10			0.98
No	4	4	4	0.713	[0.55-	2	6	4	0.243	[0.64-	5	3	4	0.176	[0.77-	5	3	4	0.439	[0.35-	6	2	4	1.000	[0.63-
					2.43]					7.59]					2.35]					1.30]					1.53]

Unknown = patients with HNSCC and no available clinicopathological characteristics. HNSCC, head and neck squamous cell carcinoma; HPV, human papillomavirus; CI, confidence interval; meth, methylated; unmeth, unmethylated; RR, relative risk.

Table IV.

Cox regression analysis of overall survival.

Variables	P-value	HR [95% CI]
DAPK1 meth	0.001^a	0.096 [0.03–0.37]
CDH1 meth	0.110	2.990 [0.78–11.48]
PAX1 meth	0.825	1.230 [0.20–7.55]
TIMP3 meth	0.951	1.040 [0.30–3.66]
CALCA meth	0.259	0.500 [0.15–1.68]
HPV-16⁻	0.006^a	9.720 [1.94–48.72]
Tumor site
Larynx	0.765	1.360 [0.18–9.98]
Oral Cavity	0.920	0.870 [0.06–13.30]
Pharynx	0.033^a	14.170 [1.25–161.15]
Tumor differentiation
Well	0.226	2.760 [0.53–14.27]
Moderate	0.013^a	8.560 [1.58–46.46]
Tumor stage
I	0.997	1.000 [0.05–19.80]
II	0.324	3.080 [0.03–28.90]
III	0.668	1.500 [0.24–9.37]
Positive lymph nodes	0.401	0.490 [0.09–2.61]
Heavy smoking	0.969	0.940 [0.05–16.71]
Heavy drinking	0.523	1.720 [0.33–9.10]
Age	0.002^a	1.110 [1.04–1.18]

Indicates a statistically significant difference (P<0.05). Meth, methylated; HPV, human papillomavirus; HR, hazard ratio; CI, confidence interval.