- Email: [email protected]
Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
CAN13454_proof ■ 19 August 2017 ■ 1/11
Cancer Letters xxx (2017) 1e11
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
Original Article
Q6 Q5
Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer Takuya Nakagawa a, b, Keisuke Matsusaka b, Kiyoshi Misawa c, Satoshi Ota d, Kiyoko Takane b, Masaki Fukuyo b, Bahityar Rahmutulla b, Ken-ichi Shinohara b, Naoki Kunii a, Daiju Sakurai a, Toyoyuki Hanazawa a, Hisahiro Matsubara e, Yukio Nakatani d, Yoshitaka Okamoto a, Atsushi Kaneda b, * a
Department of Otorhinolaryngology, Head and Neck Surgery, School of Medicine, Chiba University, Japan Department of Molecular Oncology, School of Medicine, Chiba University, Japan c Department of Otolaryngology/Head and Neck Surgery, Hamamatsu University School of Medicine, Japan d Department of Pathology, Chiba University Hospital, Japan e Department of Frontier Surgery, Graduate School of Medicine, Chiba University, Japan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 May 2017 Received in revised form 3 August 2017 Accepted 7 August 2017
Oropharyngeal squamous cell carcinoma (OPSCC) incidence has increased dramatically due to human papillomavirus (HPV); however, associated epigenetic alterations are not well studied. We performed genome-wide DNA methylation analysis using an Infinium 450k BeadArray for clinical OPSCC and noncancerous samples and cancer cell lines with/without 5-aza-20 -deoxycytidine and/or trichostatin A treatment. Frequent promoter hypermethylation and methylation-associated silencing were detected in 144 genes, which included those involved in cell-cell signaling and neuron differentiation. The methylation of nine genes (GHSR, ITGA4, RXRG, UTF1, CDH8, FAN19A4, CTNNA2, NEFH, and CASR) was quantitatively validated in 70 pharyngeal SCC cases by pyrosequencing. Hypermethylation significantly correlated with HPV-L1 positivity, but not with age or smoking status. p16INK4A was generally activated in HPV-L1(þ) tumors, and p16-positive cases significantly associated with better prognosis. RXRG hypermethylation strongly correlated with positivity of HPV-L1 and p16 (P ¼ 3 105 and P ¼ 5 104, respectively). RXRG-methylation(þ) significantly associated with better prognosis when analyzing all tumor cases (P ¼ 0.04), and when analyzing the p16-negative poorer-outcome group (P ¼ 0.03). Thus, aberrant DNA methylation might be involved in HPV-associated OPSCC; in addition, DNA methylation could serve as a marker to classify subgroups based on outcome. © 2017 Elsevier B.V. All rights reserved.
Keywords: DNA methylation Human papillomavirus (HPV) Oropharyngeal cancer Hypopharyngeal cancer
Introduction Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide; its annual incidence is 600,000 cases, accounting for 3e5% of all cancers [1]. In particular, the incidence of oropharyngeal squamous cell carcinoma (OPSCC) has increased over the last two decades, mainly because the number of human papillomavirus (HPV)-associated OPSCC cases has risen dramatically [2,3]. HPV-associated and HPV-negative OPSCCs are clinically and biologically distinct. The major risk factors for HPVnegative tumors are tobacco and excessive alcohol use. In
Q1
* Corresponding author. Department of Molecular Oncology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan. E-mail address: [email protected] (A. Kaneda).
contrast, HPV-associated tumors strongly associate with sexual behavior [4]. Compared to patients with HPV-negative tumors, those with HPV-associated tumors tend to be younger, have higher socioeconomic status, and have disease with lower primary stages and high nodal stages [5]. Regarding prognosis, it was shown that HPV(þ) status is a consistent determinant of better prognosis, regardless of treatment modality, with 5-year survival rates among patients with HPV-associated tumors of approximately 75e80% versus 45e50% among patients with HPV-negative tumors [6]. HNSCCs, especially HPV-negative tumors, are thought to initiate and progress through a series of genetic alterations caused by smoking and excessive alcohol intake. Several tumor suppressive genes are indeed dysregulated through genetic alterations, such as TP53 and CDKN2A, which are associated with the highest mutation rates, based on TCGA data for HNSCC [7]. The prevalence of TP53
http://dx.doi.org/10.1016/j.canlet.2017.08.008 0304-3835/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 2/11
2
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
and CDKN2A mutations in HNSCC ranges from 30 to 84% and 4e74%, respectively, according to various studies [8e10]. However, these two genes are not mutated in HPV-associated tumors; instead, TP53 is inactivated by the E6 oncoprotein [11e14]. HPV induces cancer by infecting epithelial basal cells. Initially after infection, HPV can be identified in circular extra chromosomal particles or episomes. Then, the viral genome typically integrates into the host cell genome, resulting in upregulated expression of viral E6 and E7 oncoproteins [10,15]. E6 and E7 proteins bind to and inactivate the tumor suppressors TP53 and RB1 respectively. This enables the host cells to avoid apoptosis and leads to proliferation [13]. In addition, cell cycle pathway feedback results in overstimulation of cell senescence pathways including p16INK4A. We can therefore use p16INK4A expression as a surrogate marker for HPV infection [16]. HPV infection is generally evaluated by detection of HPV viral oncogenes E6 and E7 using quantitative reverse transcriptase-PCR (RT-PCR) on RNA isolated from fresh frozen samples, or also by detection of L1 region of HPV by PCR combined with immunostaining of p16INK4A [16e19]. DNA methylation is one of the key epigenetic mechanisms that regulates gene expression, and aberrant DNA hypermethylation of gene promoter regions is a major inactivating mechanism to silence tumor suppressor genes in most cancers [20]. In particular, virus infection could induce aberrant DNA methylation during carcinogenesis [21,22]. Therefore, differences in methylation profiles between HPV-associated and HPV-negative OPSCCs have been explored in several studies. Whereas a limited number of similar genes have been investigated, e.g. TIMP3 and p16INK4A [23], detailed DNA methylation profiles of OPSCC have not been analyzed. Moreover, the molecular mechanisms responsible for clinical differences between patients with HPV-associated and HPV-negative tumors have not been clarified. In this study, we performed DNA methylation analysis of OPSCC on a genome-wide scale. We identified candidate methylationassociated genes by Infinium 450k methylome analysis of clinical OPSCC and non-cancerous mucosa samples, and RNA-sequencing (RNA-seq) analysis combined with Infinium 450k analysis for cancer cell lines treated with/without 5-aza-20 -deoxycytidine (Aza) and/or trichostatin A (TSA). We validated methylation status using additional clinical samples by pyrosequencing, and compared methylation status with clinicopathological factors, revealing a significant association with critical factors such as HPV status and prognosis. Materials and methods Full information is provided in the Supplementary Materials and Methods. Clinical samples and cell lines Clinical specimens were collected from 70 patients with pharyngeal squamous cell carcinoma (PSCC) who underwent therapy at Chiba University Hospital or Hamamatsu University hospital and written informed consent was obtained. Biopsy specimens were immediately frozen in liquid nitrogen and stored at 80 C. Frozen materials were microscopically examined by two independent pathologists and were dissected to enrich cancer cells when necessary. Fifty-three samples containing >50% cancer cells were used for subsequent molecular analyses. Seventeen samples were extracted from formalin-fixed paraffin-embedded (FFPE) tissues. The UMSCC47 cell line, derived from a primary tumor of the lateral tongue, was kindly provided by Dr. Thomas E. Carey, University of Michigan. The FaDu cell line, derived from primary hypopharyngeal SCC, was obtained from the ATCC. DNA was extracted by using a QIAquick DNA mini-kit (Qiagen, Hilden, Germany). The protocol was approved by the institutional review board at Chiba University and Hamamatsu University. Immunohistochemistry (IHC) IHC for TP53 was conducted using a DO-7 anti-mouse monoclonal antibody (anti-p53 DO-7 Primary Antibody, Roche, Basel, Switzerland) and the BenchMark ULTRA automated staining system (Roche). Samples with diffuse strong positive nuclear staining were considered TP53-IHC(þ) and designated as TP53-mutation(þ),
those with negative nuclear staining were considered TP53-IHC() and designed as TP53-mutation(þ), and those with sporadic positive nuclear staining were considered TP53-IHC wild-type and designed as TP53-mutation(). For p16INK4A, IHC was performed using an anti-mouse monoclonal antibody (CINtec p16 Histology, Roche) and BenchMark ULTRA, and samples with positive nuclear staining (nuclear and cytoplasmic immunoreactivity in >70% of tumor cells) were considered p16-IHC(þ) and designated as p16(þ). Amplification of L1 DNA region in high risk HPV Evaluation of HPV infection was also performed by amplifying a portion of the L1 region in high risk HPV using GP5 and GP6 primers, as previously reported [24]. Samples with the PCR amplicon was designated as HPV-L1(þ). Aza and TSA treatment UMSCC47 and FaDu cells were seeded at a density of 3 105 cells/10-cm dish on day 0. Cells were exposed to Aza (3 mM for UMSCC47 and 5 mM for FaDu) or vehicle (DMSO) on days 1, 2, and 3. Cells were exposed to TSA (300 nM for UMSCC47 and 100 nM for FaDu) or vehicle (ethanol) on day 3. Medium was changed every 24 h, and cells were harvested on day 4. Infinium assays The Infinium HumanMethylation450 BeadChip (Illumina) contains approximately 485,000 individual CpG sites covering 99% of RefSeq genes with an average of 17 CpG sites per gene. For each CpG site, b-value, ranging from 0.00 to 1.00, was measured by a methylated probe relative to the sum of both methylated and unmethylated probes. Bisulfite conversion was performed using the Zymo EZ DNA Methylation Kit (Zymo Research, Irvine, CA) with 500 ng of genomic DNA for each sample. Whole genome amplification, labeling, hybridization, and scanning were performed according to the manufacturer's protocols. Infinium analysis was performed against 13 clinical OPSCC, four non-cancerous mucosa samples, and cancer cell lines treated with/without Aza and/or TSA (Fig. 1A), and Infinium data were submitted to the GEO database under the accession numbers GSM2612419eGSM2612439. For analysis of DNA methylation at promoter regions based on Infinium data, the probe nearest to the transcription start site (TSS) was selected when multiple probes were designed for one promoter region. The CpG score for each probe was calculated based on a previous report [25,26]. For analysis of CpG island shore regions upstream and downstream from TSS, the probe nearest to TSS was selected among probes at CpG island shore (annotated by Illumina). For analysis of gene body, the probe showing the highest CpG score was selected among probes at gene body, i.e. þ4 kb downstream from TSS through the transcription terminal site. RNA-seq analysis Libraries for RNA-seq were prepared using the TruSeq Stranded mRNA Sample Prep Kit (Illumina, San Diego, CA), following the manufacturer's protocol. Deep sequencing was performed on the Illumina HiSeq 1500 platform using the TruSeq Rapid SBS Kit (Illumina) in 50-base single-end mode according to the manufacturer's protocol. RNA-seq analysis was performed against cancer cell lines treated with/ without Aza and/or TSA (Fig. 1A), and RNA-seq data were submitted to the GEO database under the accession numbers GSM2612408 e GSM2612411. Gene Ontology (GO) analysis Gene annotation enrichment analysis was conducted based on GO (biologic process, cellular component, and molecular function) using the Functional Annotation tool at DAVID Bioinformatics Resources (http://david.abcc.ncifcrf.gov/). Pyrosequencing analysis Quantitative validation of methylation in 70 cases was performed by pyrosequencing using the PyroMark Q96 (Qiagen) as previously reported [26]. Primers were designed by Pyro Q-CpG Software (Qiagen) to amplify bisulfite-treated DNA (Supplementary Table S1). Methylation control samples (0%, 25%, 50%, 75%, and 100%) were prepared as described previously [27] and used to confirm the high quantification ability of pyrosequencing assays. Statistical analysis Association between clinicopathological factors and HPV-L1 status or DNA methylation was analyzed using the Fisher's exact test, c2 test, Student's t-test, ANOVA, and Kolmogorov-Smirnov test. Unsupervised two-way hierarchical clustering was performed using Cluster 3.0 software. Overall Survival (OS) was measured from registration date until date of PSCC-related death, filtering patients alive at last follow-up and non-PSCC-related deaths. Estimation of OS distribution was performed by the Kaplan-Meier method. Differences between groups were determined by a log-rank test using R software (www.r-project.org/). P < 0.05 was considered statistically significant.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 3/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
3
Fig. 1. The design of the study, immunostaining, and HPV status in OPSCC. (A) A flowchart showing the design of this study. (B) Immunostaining of p16 and TP53. Sections of OPSCC tissues were stained with hematoxylin and eosin (left), or analyzed by immunostaining for p16 (middle) and TP53 (right). TP53 was wild type in most p16(þ) OPSCC samples (Case #1). Most of p16() OPSCC cases were TP53-mutation(þ), showing either positive staining of TP53 (Case #2) or negative staining of TP53 (Case #3). Scale bar, 100 mm. (C) HPV infection status. The presence of a 142-bp PCR product for HPV-L1 region was analyzed by gel electrophoresis. Most p16(þ) cases were HPV-L1(þ) (Case #1), whereas p16() cases were generally HPV-L1().
Results Immunostaining analyses and HPV status Expression of p16INK4A and TP53 in HPV-associated OPSCC, HPVnegative OPSCC, and HPSCC was analyzed by IHC (Fig. 1B). HPV
status was evaluated by PCR amplification of L1 region in high risk HPV with GP5/GP6 primers (Fig. 1C). All 22 HPV-associated OPSCC samples that underwent IHC were p16(þ), whereas only one HPVnegative OPSCC specimen was p16(þ). Most p16(þ) cases, except two, were TP53 wild-type.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 4/11
4
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
DNA methylation analysis using Infinium 450k and RNA-seq analysis We performed Infinium 450k BeadArray analysis on 13 clinical OPSCC and four non-cancerous mucosa samples (Fig. 2A). To analyze Infinium data, the probe nearest the TSS was selected for each gene, and b-values were compared by a t-test. A total of 2093 probes in high-CpG promoters had a P-value < 0.05, and these were used for hierarchical clustering. Unsupervised 2-way hierarchical clustering analysis of 13 OPSCC and four normal mucosa samples classified these samples into two clusters. The low methylation cluster included the lowest methylation subgroup, consisting of the four non-cancerous mucosa samples, and a relatively low methylation subgroup of three OPSCC samples. The high methylation cluster included 10 OPSCC cases divided into two subgroups. Interestingly, all five HPV-L1(þ) OPSCC samples were included in the high methylation cluster, forming a unique subcluster, whereas OPSCC samples with relatively low methylation were all HPV-L1(). When a cluster of 487 genes were extracted from the 2093 high-CpG genes, OPSCC samples showed markedly higher methylation levels than non-cancerous samples (0.34 ± 0.00 vs 0.16 ± 0.00, P < 1 1030, t-test) (Fig. 2A). We also investigated CpG island shore regions upstream and downstream from TSS, and gene body regions. The upstream CpG island shore regions were significantly hypermethylated in cancer samples than non-cancerous samples (0.43 ± 0.00 vs 0.35 ± 0.01, P ¼ 3 1030). These regions were already hypermethylated to some extent in non-cancerous mucosa, and the methylation level did not increase in cancer so markedly as observed in high-CpG promoter regions. The downstream CpG island shore was generally hypermethylated in non-cancerous samples, and methylation levels in cancer were not significantly different from non-cancerous samples (0.54 ± 0.00 vs 0.55 ± 0.00, P ¼ 0.2). Gene body showed aberrant hypomethylation in cancer; methylation levels were decreased from 0.72 ± 0.00 in non-cancerous mucosa to 0.61 ± 0.00 in cancer (P < 1 1030) (Fig. 2A). Since methylation levels altered most apparently in high-CpG promoter regions, we focused on these regions for the subsequent analyses. We next extracted frequently hypermethylated and silenced genes in OPSCC samples. Among the 13,221 high-CpG genes with an average b-value <0.2 in the four non-cancerous mucosa samples, 625 were hypermethylated (b-value >0.3) frequently in 3 of 13 OPSCC samples (Fig. 2B). We next performed RNA-seq analysis on UMSCC47 and FaDu cell lines before and after Aza and/or TSA treatment, and found 3595 and 2778 genes showing >2-fold upregulation in UMSCC47 and FaDu cell lines, respectively, compared to expression levels in untreated cells. Among the 625 frequently hypermethylated genes, 144 were highly methylated (bvalue >0.6) in at least one cell line before Aza/TSA treatment, with a >2-fold upregulation after Aza/TSA treatment, accompanying a decrease in the b-value (Fig. 2B). For these 144 genes, we confirmed the unmethylated status in all four non-cancerous mucosa samples and the hypermethylated status in OPSCC cases in the surrounding probes as well as the selected probe nearest to the TSS (Fig. 2C). GO terms related to cellcell signaling and neuron differentiation were significantly enriched (Fig. 2D). Validation of DNA methylation by pyrosequencing Nine genes were chosen from the 144 frequently silenced genes, including those related to relevant GO terms, and methylation levels were validated by quantitative methylation analysis by pyrosequencing. First, we analyzed the methylation level at the
selected probe site using OPSCC and non-cancerous mucosa samples that were previously analyzed by Infinium (Fig. 3A). Methylation levels were confirmed to be highly similar. p16INK4A was additionally analyzed, and confirmed to be unmethylated in both non-cancerous mucosa and OPSCC samples. We next investigated the methylation status of these nine genes in pharyngeal cancer samples including 42 OPSCC and 15 hypo-pharyngeal SCC by pyrosequencing (Fig. 3B). Not only the selected probe site, but also the surrounding CpG sites were similarly hypermethylated, indicating that aberrant DNA methylation does not occur randomly at a single CpG site, but occurs at multiple consecutive CpG sites in a region near the TSS. When tumor samples were ordered (from the left) based on average methylation level, HPV-L1(þ) samples were significantly enriched on the left (P ¼ 0.01, Kolmogorov-Smirnov test). This indicated that HPV-positivity significantly correlates with frequent DNA hypermethylation. However, other clinicopathological factors such as age or smoking status were not significantly correlated with DNA methylation level. When DNA methylation levels were compared between HPV-L1(þ) and HPVL1() samples, significantly higher methylation was detected in all the cancer samples (P ¼ 0.0002, t-test) or in OPSCC samples (P ¼ 0.001, t-test). Comparison of HPV status, IHC data, and methylation status As p16INK4A is a surrogate marker of HPV infection, p16(þ) status correlated strongly with HPV-L1(þ). TP53-mutation(þ) cases included 21 cases with nuclear staining and 16 cases with negative staining, and these were strongly, inversely correlated with p16(þ) cases (Fig. 4A). To compare methylation status with HPV status and IHC data, we investigated the methylation status of the four most frequently methylated genes (RXRG, GHSR, CTNNA2, and ITGA4). Aberrant methylation was significantly correlated with HPV-L1(þ) cases, and RXRG methylation showed a particularly strong correlation with HPV-positivity (P ¼ 3 105, c2-test) (Fig. 4B). Despite the significant correlation, not only p16(þ) cases, but also some p16() cases showed aberrant RXRG methylation. We also analyzed the association between clinicopathological factors and HPV-L1 status, or RXRG methylation levels (Table 1). HPV-L1 positivity associated significantly with tumor location (tonsils), p16(þ), wild-type TP53, and alcohol() (P ¼ 0.002, 2 1012, 3 108, 0.002, respectively). These associations were in good agreement with the previous reports that HPV-associated OPSCC arises within tonsils and base of tongue, that HPVassociated OPSCC patients do not intake excessive alcohol, and that TP53 mutation is generally found in HPV-negative cases, not in HPV-associated cases [28]. High RXRG methylation levels also associated significantly with tumor location (tonsils), p16(þ), and wild-type TP53 (P ¼ 0.04, 5 104, 0.002, respectively). We then investigated the association with prognosis. p16(þ) cases and L1(þ) cases, which are suggested to be HPV-associated tumors, significantly associated with better prognosis than p16() cases (P ¼ 0.002, log-rank test) and L1() cases (P ¼ 0.005), respectively (Fig. 5A). Whereas TP53-mutation did not predict outcome, RXRG-methylation(þ) cases significantly associated with better prognosis when analyzing all cancer cases (P ¼ 0.04) (Fig. 5A). When analyzing a p16(þ) better-outcome group and a p16() poorer-outcome group separately, all the p16(þ) cases showed good prognosis regardless of RXRG methylation status. Among p16() poorer-outcome patients, RXRG-methylation(þ) correlated significantly with good prognosis (P ¼ 0.03) (Fig. 5B). From these results, two-step panel might be generated to predict prognosis. The first panel “p16 IHC” extracts p16(þ) better-outcome patients, and the second panel “RXRG methylation” additionally
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 5/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
5
Fig. 2. Infinium 450k analysis of OPSCC and non-cancerous samples. (A) Genome-wide DNA methylation analyzed for 13 clinical OPSCC and four non-cancerous samples. A probe nearest to the transcriptional start site (TSS) was selected for analysis of each promoter region. Probes with significantly different b-values between OPSCC and non-cancerous samples (P < 0.05, t-test) were extracted, and 2093 genes with high-CpG promoters (CpG score >0.48) were used for clustering (top). Genes with low-CpG promoters (CpG score <0.48), CpG island shore regions upstream and downstream from TSS, and gene body regions are additionally shown (middle). White: HPV-L1(), age<60, smoking(); black: HPV-L1(þ), age>60, smoking(þ) (bottom). Whereas three OPSCC samples with relatively low methylation formed a cluster with non-cancerous samples, other OPSCC samples comprised a cluster with high methylation. Methylation levels between tumor and non-cancerous samples were compared using t-test (right panels). The difference of methylation levels was the most apparent in high-CpG promoter region (P < 1 1030, t-test). (B) There were 144 genes that met the following criteria: (i) unmethylated in four non-cancerous samples (average bvalue <0.2); (ii) frequently hypermethylated in 3 of 13 clinical OPSCC samples (b-value >0.3); (iii) hypermethylated in at least one cancer cell line before Aza/TSA treatment (b-value >0.6); (iv) de-methylated after Aza/TSA treatment (i.e. showing a decreased of b-value). (C) For the 144 frequently silenced genes, we confirmed that they were unmethylated in the four non-cancerous mucosa samples and hypermethylated in OPSCC cases in the surrounding probes as well as the selected probe nearest to the TSS. Representatively, GHSR was unmethylated in four non-cancerous samples (N), but OPSCC samples (T) were hypermethylated at the probe site nearest to the TSS and the surrounding probe sites as well. Arrowhead, the position of TSS. (D) For these genes, Gene Ontology (GO) terms related to cell-cell signaling and neuron differentiation were significantly enriched.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 6/11
6
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
Fig. 3. Validation of DNA methylation in OPSCC samples by pyrosequencing. (A) Nine genes were chosen from the 144 frequently silenced genes. The methylation levels (analyzed by Infinium) are shown in the color scale (top), and the methylation levels at the probe site was also analyzed by pyrosequencing and shown (bottom), confirming similar results. p16 was additionally analyzed. (B) Pyrosequencing data of 70 pharyngeal cancer samples. To investigate the methylation status in other pharyngeal cancer samples and the relationship with clinicopathological factors (bottom), methylation levels in 42 OPSCC and 15 hypo-pharyngeal SCC samples were additionally analyzed by pyrosequencing (top). Closed boxes: HPV-L1(þ), oropharyngeal, elder age >60 years old, smoking(þ), respectively. Tumor samples were ordered from left to right based on average methylation level of the nine genes; HPV-L1(þ) samples were significantly enriched on the left (P ¼ 0.01, Kolmogorov-Smirnov test), i.e. HPV infection correlated significantly with frequent hypermethylation. HPVL1(þ) samples were associated with higher methylation levels within all tumor samples (P ¼ 0.0002, t-test) and also within OPSCC (P ¼ 0.001, t-test). Frequent hypermethylation did not correlate with age or smoking status. Whereas red numbers indicated the CpG sites of the analyzed Infinium probes, the surrounding CpG sites as well as the probe site were shown to be hypermethylated at similar levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 7/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
7
Fig. 4. Comparison of HPV status, immunostaining data, and methylation status in OPSCC samples. (A) Left, HPV-L1 status and IHC data of p16 and TP53 (See Fig. 1 for representative results). Right, the methylation status of the four most frequently methylated genes (RXRG, GHSR, CTNNA2, and ITGA4). White, methylation level <50%. Black, methylation level 50%. HPV-L1(þ) status correlated strongly with p16(þ) status (P < 1 1015, Fisher's exact test), and TP53-mutation(þ) cases showed a strong inverse correlation with p16(þ) cases (P ¼ 2 1010). (B) Correlation between aberrant methylation and HPV-L1(þ) status. For each of the four genes, especially RXRG, aberrant methylation significantly correlated with HPV-L1(þ) samples.
extracts RXRG-methylation(þ) better-outcome patients among p16() group (Fig. 5C). Discussion The involvement of epigenetic alterations in OPSCC has not been clarified; thus, we analyzed aberrant DNA methylation in OPSCC on
a genome-wide scale using the Infinium 450k BeadArray. We identified frequently hypermethylated and silenced genes in OPSCC, such as RXRG, and showed that these genes are preferentially hypermethylated in HPV-positive tumors. p16INK4A is activated in HPV-positive tumors, and p16(þ) cases show significantly better prognosis than p16() cases. Moreover, RXRG-methylation(þ) cases significantly associated with better prognosis when
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 8/11
8
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
Table 1 Clinicopathological factors and their association with DNA methylation. All cases
# of cases 70 Gender Male 60 Female 10 Age >60 51 <60 19 Tumor location tonsils 34 base of tongue 8 soft palate 11 posterior wall 2 Clinical tumor classification T1 6 T2 33 T3 16 T4a 14 T4b 1 Clinical lymph node classification N0 23 N1 4 N2a 3 N2b 25 N2c 12 N3 3 Lymph node metastasis (þ) 47 () 23 Clinical stage I 6 II 11 III 8 IVA 39 IV B 6 p16 IHC (þ) 23 () 39 N/A 8 TP53 IHC wild type 25 mutation(þ) 37 N/A 8 Smoking status (þ) 51 () 9 N/A 10 Alcohol status (þ) 47 () 12 N/A 11
HPSCC
OPSCC
RXRG methylation level
L1(þ)
L1()
P-value
Methylation (%)
P-value
15
24
31
12 3
20 4
28 3
0.7
36.3 ± 3.1 46.9 ± 10.2
0.4
9 6
18 6
24 7
1
38.4 ± 3.6 37.5 ± 6.0
0.9
e e e e
21 3 0 0
13 5 11 2
0.002*
47.3 35.8 30.0 44.4
± ± ± ±
4.1 9.5 4.2 41.6
0 3 5 7 0
0 16 5 2 1
6 14 6 5 0
± ± ± ±
8.2 4.8 6.3 6.6
0.2
32.9 43.5 32.6 34.2 55
3 2 0 6 4 0
6 1 3 9 3 2
14 1 0 10 5 1
0.3
30.4 30.2 62.8 40.0 41.0 57.8
± ± ± ± ± ±
12 3
18 6
17 14
0.2
41.8 ± 3.6 30.4 ± 5.5
0 0 4 11 0
0 5 2 12 5
6 6 2 16 1
0.07
32.9 30.6 30.3 39.9 58.0
0 14 1
23 0 1
1 24 6
2 10e12*
52.9 ± 5.2 29.4 ± 3.5
5 10e4*
4 10 1
20 2 2
3 23 5
3 10e8*
50.7 ± 5.3 29.6 ± 3.5
0.002*
8 1 6
17 6 1
26 2 3
0.1
41.3 ± 3.4 43.4 ± 10.4
0.9
9 0 6
12 10 2
26 2 3
0.002*
39.8 ± 3.7 50.7 ± 7.0
0.2
0.04*
0.7
± ± ± ± ±
5.5 13.7 13.9 4.6 8.3 5.0
8.2 10.5 8.8 4.1 2.0
0.09
0.1
0.09
OPSCC, oropharyngeal squamous cell carcinoma. HPSCC, hypo-pharyngeal squamous cell carcinoma. N/A, not applicable. For OPSCC, association with HPV-L1 status was analyzed by Fisher's exact test or c2 test. For RXRG methylation, methylation levels were compared by t-test or ANOVA. *P < 0.05.
analyzing all tumor cases, and also when analyzing the p16() poorer-outcome group. More than 10 genes were previously reported to be inactivated by DNA methylation in OPSCC [11,29,30], including p16INK4A, HOXA9, MGMT, GATA4, and RASSF1 [23,30]. Herein, 144 genes were extracted as frequently hypermethylated and silenced in OPSCC, but p16INK4A was rarely methylated and was not included. GO terms related to neuron differentiation and cell development were enriched in these genes, in agreement with previous studies reporting that genes related to differentiation and development are targets of the polycomb complex in ES cells and that these genes are preferentially hypermethylated in many types of cancer [26,31]. Lechner et al. reported 43 hypermethylated promoter regions associated with HPV in HNSCC through Infinium analysis, including three cadherins of the Polycomb group target genes [32]. Weiss
et al. investigated promoter methylation of 12 genes (TIMP3, CDH1, p16INK4A, DAPK1, TCF21, CD44, MLH1, MGMT, RASSF1, CCNA1, LARS2, and CEPBA) by methylation-specific PCR, which is not a quantitative method, in a cohort of 55 HNSCC cases derived mostly from the oropharynx. HPV-associated tumors significantly correlated with CCNA1 and TIMP3 promoter methylation [23]. Schlecht et al. reported 22 CpG loci showing significant difference of methylation level between HPV-positive and HPV-negative OPSCC, which included four CpG loci at CDKN2A downstream region, but not promoter region around TSS [33]. O'Regan et al. showed frequent p16INK4A hypermethylation in HNSCC of anterior two thirds of the tongue, but no p16INK4A methylation in OPSCC [34]. The association of HPV-positive OPSCC with DNA hypermethylation of promoter regions, including p16 promoters, is thus still controversial. In this study, a highly quantitative analysis based on pyrosequencing
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 9/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
9
Fig. 5. Analysis of prognosis in pharyngeal cancer cases. (A) Analysis of prognosis based on p16, HPV-L1, TP53 mutation, and RXRG methylation status. p16(þ) cases were significantly associated with better prognosis than p16() cases as reported (P ¼ 0.002, log-rank test), and HPV-L1(þ) cases that were mostly p16(þ) were also significantly correlated with better outcome (P ¼ 0.005). Whereas TP53-mutation did not predict outcome, RXRG-methylation(þ) cases were significantly associated with better prognosis when analyzing all cancer cases (P ¼ 0.04). (B) RXRG methylation correlated with better outcome even when analyzing the p16() poorer-outcome group (P ¼ 0.03). p16(þ) cases showed good prognosis regardless of RXRG methylation status. (C) Two-step panel to predict the prognosis of OPSCC. The first and second panels classify patients into a good prognosis group with p16(þ) and/or RXRG-methylation(þ), and a poor prognosis group with p16() and RXRG-methylation().
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 10/11
10
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
revealed that hypermethylation of p16INK4A promoter region is very rare. Instead, promoter regions of 144 genes were found to be frequently hypermethylated in OPSCC, and analysis of nine genes showed that frequent DNA methylation is significantly correlated with HPV-positive status. Other than promoter CpG islands, CpG island shore and gene body regions were also analyzed. Gene body generally showed high methylation levels in non-cancerous mucosa, and significant decrease of methylation levels was observed in OPSCC. This is consistent with global DNA hypomethylation reported in many types of cancer [35]. CpG island shore regions upstream from TSS were hypermethylated in OPSCC, but the difference of methylation levels between cancer and non-cancerous samples was smaller than promoter regions. Aberrant DNA methylation is reported to occur from lower-CpG peripheral regions to higher-CpG regions in high-CpG promoters in carcinogenesis, and aberrant methylation at CpG island shore is frequently observed in the surrounding noncancerous tissues [36,37]. The expression of p16INK4A is used as a surrogate marker of HPV infection. E6 and E7 oncoproteins are known to inactivate TP53 and RB1, which enables the host cells to avoid apoptosis and proliferate; this simultaneously overstimulates cellular senescence pathways because of cell cycle pathway feedback. p16INK4A is thus strongly expressed in HPV-positive OPSCC. Our data also indicated that hypermethylation of p16INK4A is not the cause of p16() status in HPV-negative cases. It has been reported that HPV-negative HNSCC featured novel focal deletions in CDKN2A [28]. Further analysis is necessary to clarify whether negative expression of p16 is due to gene mutation or other causes. Interestingly, although there was a significant correlation between DNA hypermethylation and HPVpositivity, a fraction of HPV-negative OPSCC also exhibited DNA hypermethylation. Whereas p16 positivity (thus HPV positivity) is a marker of better OPSCC prognosis [5,6], RXRG-methylation(þ) cases showed better outcome among all tumor cases, and even among the p16() poorer-outcome group. Retinoid X receptors (RXRs) are nuclear receptors for retinoids that have important roles in the regulation of growth and differentiation. RXRs have three separate isotypes encoded on three separate genes: RXRA, RXRB, and RXRG. Promoter hypermethylation of RXRG and the loss of gene expression were reported in lung cancer [38]. Although whether RXRG methylation plays a critical role in carcinogenesis is yet to be investigated, we identified frequent hypermethylation of RXRG in OPSCC and its significant correlation with better prognosis. Further studies are necessary to clarify how HPV infection is associated with DNA hypermethylation, or alternatively, whether there is another independent mechanism that confers relatively better outcome in OPSCC with higher DNA methylation. It is also reported that exposure to selective RXRG ligands decrease cellular proliferation of thyroid cancer cell lines expressing RXRG [39,40]. Although further studies should be investigated, RXRG might be a potential target for treatment of RXRG-expressing cancer [41]. In summary, in OPSCC, we identified frequently hypermethylated and silenced genes that are preferentially hypermethylated in HPV-associated tumors. While p16INK4A expression is known to be a marker of better prognosis in OPSCC, RXRG methylation(þ) cases show better prognosis among all the cases as well as within p16() cases. Acknowledgements We thank Chihomi Sato, Keiko Iida, and Aki Komatsu for technical assistance, and Editage (www.editage.jp) for English language editing. This study was supported by Japan Agency for Medical Research and Development (AMED) [Practical Research for
Innovative Cancer Control 17ck0106263h0001 to AK]; Japan Society for the Promotion of Science (JSPS) [24249081 to YO, 16H05412 to AK]; and a grant from Global and Prominent Research, Chiba University to AK. Q2 Conflict of interest All authors have no conflicts of interest to disclose. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.canlet.2017.08.008. References [1] J. Ferlay, H.R. Shin, F. Bray, D. Forman, C. Mathers, D.M. Parkin, Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008, Int. J. Cancer 127 (2010) 2893e2917. [2] A.K. Chaturvedi, E.A. Engels, W.F. Anderson, M.L. Gillison, Incidence trends for human papillomavirus-related and -unrelated oral squamous cell carcinomas in the United States, J. Clin. Oncol. 26 (2008) 612e619. [3] A.B. Ryerson, E.S. Peters, S.S. Coughlin, V.W. Chen, M.L. Gillison, M.E. Reichman, et al., Burden of potentially human papillomavirus-associated cancers of the oropharynx and oral cavity in the US, 1998-2003, Cancer 113 (2008) 2901e2909. [4] M.L. Gillison, G. D'Souza, W. Westra, E. Sugar, W. Xiao, S. Begum, et al., Distinct risk factor profiles for human papillomavirus type 16-positive and human papillomavirus type 16-negative head and neck cancers, J. Natl. Cancer Inst. 100 (2008) 407e420. [5] K. Grisar, R. Dok, J. Schoenaers, T. Dormaar, E. Hauben, M. Jorissen, et al., Differences in human papillomavirus-positive and -negative head and neck cancers in Belgium: an 8-year retrospective, comparative study, Oral Surg. Oral Med. Oral Pathol. Oral Radiol. 121 (2016) 456e460. [6] K.K. Ang, J. Harris, R. Wheeler, R. Weber, D.I. Rosenthal, P.F. Nguyen-Tan, et al., Human papillomavirus and survival of patients with oropharyngeal cancer, N. Engl. J. Med. 363 (2010) 24e35. [7] N. Agrawal, M.J. Frederick, C.R. Pickering, C. Bettegowda, K. Chang, R.J. Li, et al., Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1, Science 333 (2011) 1154e1157. [8] H. Blons, P. Laurent-Puig, TP53 and head and neck neoplasms, Hum. Mutat. 21 (2003) 252e257. [9] D.N. Hayes, C. Van Waes, T.Y. Seiwert, Genetic landscape of human papillomavirus-associated head and neck cancer and comparison to tobaccorelated tumors, J. Clin. Oncol. 33 (2015) 3227e3234. [10] W. Zhang, A. Edwards, Z. Fang, E.K. Flemington, K. Zhang, Integrative genomics and transcriptomics analysis reveals potential mechanisms for favorable prognosis of patients with HPV-positive head and neck carcinomas, Sci. Rep. 6 (2016) 24927. [11] P.M. van Kempen, R. Noorlag, W.W. Braunius, I. Stegeman, S.M. Willems, W. Grolman, Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: a systematic review, Epigenetics 9 (2014) 194e203. [12] N. Stransky, A.M. Egloff, A.D. Tward, A.D. Kostic, K. Cibulskis, A. Sivachenko, et al., The mutational landscape of head and neck squamous cell carcinoma, Science 333 (2011) 1157e1160. [13] C.H. Chung, M.L. Gillison, Human papillomavirus in head and neck cancer: its role in pathogenesis and clinical implications, Clin. Cancer Res. 15 (2009) 6758e6762. [14] M. Lechner, G.M. Frampton, T. Fenton, A. Feber, G. Palmer, A. Jay, et al., Targeted next-generation sequencing of head and neck squamous cell carcinoma identifies novel genetic alterations in HPVþ and HPV tumors, Genome Med. 5 (2013) 49. [15] M. Parfenov, C.S. Pedamallu, N. Gehlenborg, S.S. Freeman, L. Danilova, C.A. Bristow, et al., Characterization of HPV and host genome interactions in primary head and neck cancers, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) 15544e15549. [16] C. Ndiaye, M. Mena, L. Alemany, M. Arbyn, X. Castellsague, L. Laporte, et al., HPV DNA, E6/E7 mRNA, and p16INK4a detection in head and neck cancers: a systematic review and meta-analysis, Lancet Oncol. 15 (2014) 1319e1331. [17] M.M. Rietbergen, C.R. Leemans, E. Bloemena, D.A. Heideman, B.J. Braakhuis, A.T. Hesselink, et al., Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm, Int. J. Cancer 132 (2013) 1565e1571. [18] M.A. Broglie, W. Jochum, A. Michel, T. Waterboer, D. Foerbs, R. Schoenegg, et al., Evaluation of type-specific antibodies to high risk-human papillomavirus (HPV) proteins in patients with oropharyngeal cancer, Oral Oncol. 70 (2017) 43e50. [19] S.J. Smeets, A.T. Hesselink, E.J. Speel, A. Haesevoets, P.J. Snijders, M. Pawlita, et al., A novel algorithm for reliable detection of human papillomavirus in
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
Q3
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
CAN13454_proof ■ 19 August 2017 ■ 11/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28] [29]
[30]
paraffin embedded head and neck cancer specimen, Int. J. Cancer 121 (2007) 2465e2472. D.J. Weisenberger, Characterizing DNA methylation alterations from the cancer genome atlas, J. Clin. Invest. 124 (2014) 17e23. A. Kaneda, K. Matsusaka, H. Aburatani, M. Fukayama, Epstein-Barr virus infection as an epigenetic driver of tumorigenesis, Cancer Res. 72 (2012) 3445e3450. Z. Herceg, A. Paliwal, Epigenetic mechanisms in hepatocellular carcinoma: how environmental factors influence the epigenome, Mutat. Res. 727 (2011) 55e61. D. Weiss, T. Basel, F. Sachse, A. Braeuninger, C. Rudack, Promoter methylation of cyclin A1 is associated with human papillomavirus 16 induced head and neck squamous cell carcinoma independently of p53 mutation, Mol. Carcinog. 50 (2011) 680e688. K. Husnjak, M. Grce, L. Magdic, K. Pavelic, Comparison of five different polymerase chain reaction methods for detection of human papillomavirus in cervical cell specimens, J. Virol. Methods 88 (2000) 125e134. M. Weber, I. Hellmann, M.B. Stadler, L. Ramos, S. Paabo, M. Rebhan, et al., Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome, Nat. Genet. 39 (2007) 457e466. K. Matsusaka, A. Kaneda, G. Nagae, T. Ushiku, Y. Kikuchi, R. Hino, et al., Classification of Epstein-Barr virus-positive gastric cancers by definition of DNA methylation epigenotypes, Cancer Res. 71 (2011) 7187e7197. K. Yagi, K. Akagi, H. Hayashi, G. Nagae, S. Tsuji, T. Isagawa, et al., Three DNA methylation epigenotypes in human colorectal cancer, Clin. Cancer Res. 16 (2010) 21e33. N. Cancer Genome Atlas, Comprehensive genomic characterization of head and neck squamous cell carcinomas, Nature 517 (2015) 576e582. R. Guerrero-Preston, C. Michailidi, L. Marchionni, C.R. Pickering, M.J. Frederick, J.N. Myers, et al., Key tumor suppressor genes inactivated by “greater promoter” methylation and somatic mutations in head and neck cancer, Epigenetics 9 (2014) 1031e1046. N.V. Anayannis, N.F. Schlecht, T.J. Belbin, Epigenetic mechanisms of human papillomavirus-associated head and neck cancer, Arch. Pathol. Lab. Med. 139 (2015) 1373e1378.
11
[31] M. Widschwendter, H. Fiegl, D. Egle, E. Mueller-Holzner, G. Spizzo, C. Marth, et al., Epigenetic stem cell signature in cancer, Nat. Genet. 39 (2007) 157e158. [32] M. Lechner, T. Fenton, J. West, G. Wilson, A. Feber, S. Henderson, et al., Identification and functional validation of HPV-mediated hypermethylation in head and neck squamous cell carcinoma, Genome Med. 5 (2013) 15. [33] N.F. Schlecht, M. Ben-Dayan, N. Anayannis, R.A. Lleras, C. Thomas, Y. Wang, et al., Epigenetic changes in the CDKN2A locus are associated with differential expression of P16INK4A and P14ARF in HPV-positive oropharyngeal squamous cell carcinoma, Cancer Med. 4 (2015) 342e353. [34] E.M. O'Regan, M.E. Toner, S.P. Finn, C.Y. Fan, M. Ring, B. Hagmar, et al., p16(INK4A) genetic and epigenetic profiles differ in relation to age and site in head and neck squamous cell carcinomas, Hum. Pathol. 39 (2008) 452e458. [35] M. Ehrlich, DNA hypomethylation in cancer cells, Epigenomics 1 (2009) 239e259. [36] K. Matsusaka, S. Funata, M. Fukuyo, Y. Seto, H. Aburatani, M. Fukayama, et al., Epstein-Barr virus infection induces genome-wide de novo DNA methylation in non-neoplastic gastric epithelial cells, J. Pathol. (2017). Q4 [37] R.A. Irizarry, C. Ladd-Acosta, B. Wen, Z. Wu, C. Montano, P. Onyango, et al., The human colon cancer methylome shows similar hypo- and hypermethylation at conserved tissue-specific CpG island shores, Nat. Genet. 41 (2009) 178e186. [38] S.M. Lee, J.Y. Lee, J.E. Choi, S.Y. Lee, J.Y. Park, D.S. Kim, Epigenetic inactivation of retinoid X receptor genes in non-small cell lung cancer and the relationship with clinicopathologic features, Cancer Genet. Cytogenet 197 (2010) 39e45. [39] J. Klopper, M. Kane, A. Jimeno, S. Sams, J. French, L. Pike, et al., A phase II trial of bexarotene for advanced differentiated thyroid cancer, Thyroid 25 (2015) 563e564. [40] J.P. Klopper, W.R. Hays, V. Sharma, M.A. Baumbusch, J.M. Hershman, B.R. Haugen, Retinoid X receptor-gamma and peroxisome proliferatoractivated receptor-gamma expression predicts thyroid carcinoma cell response to retinoid and thiazolidinedione treatment, Mol. Cancer Ther. 3 (2004) 1011e1020. [41] B.R. Haugen, L.L. Larson, U. Pugazhenthi, W.R. Hays, J.P. Klopper, C.A. Kramer, et al., Retinoic acid and retinoid X receptors are differentially expressed in thyroid cancer and thyroid carcinoma cell lines and predict response to treatment with retinoids, J. Clin. Endocrinol. Metab. 89 (2004) 272e280.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58
CAN13454_proof ■ 19 August 2017 ■ 1/11
Cancer Letters xxx (2017) 1e11
Contents lists available at ScienceDirect
Cancer Letters journal homepage: www.elsevier.com/locate/canlet
Original Article
Q6 Q5
Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer Takuya Nakagawa a, b, Keisuke Matsusaka b, Kiyoshi Misawa c, Satoshi Ota d, Kiyoko Takane b, Masaki Fukuyo b, Bahityar Rahmutulla b, Ken-ichi Shinohara b, Naoki Kunii a, Daiju Sakurai a, Toyoyuki Hanazawa a, Hisahiro Matsubara e, Yukio Nakatani d, Yoshitaka Okamoto a, Atsushi Kaneda b, * a
Department of Otorhinolaryngology, Head and Neck Surgery, School of Medicine, Chiba University, Japan Department of Molecular Oncology, School of Medicine, Chiba University, Japan c Department of Otolaryngology/Head and Neck Surgery, Hamamatsu University School of Medicine, Japan d Department of Pathology, Chiba University Hospital, Japan e Department of Frontier Surgery, Graduate School of Medicine, Chiba University, Japan b
a r t i c l e i n f o
a b s t r a c t
Article history: Received 13 May 2017 Received in revised form 3 August 2017 Accepted 7 August 2017
Oropharyngeal squamous cell carcinoma (OPSCC) incidence has increased dramatically due to human papillomavirus (HPV); however, associated epigenetic alterations are not well studied. We performed genome-wide DNA methylation analysis using an Infinium 450k BeadArray for clinical OPSCC and noncancerous samples and cancer cell lines with/without 5-aza-20 -deoxycytidine and/or trichostatin A treatment. Frequent promoter hypermethylation and methylation-associated silencing were detected in 144 genes, which included those involved in cell-cell signaling and neuron differentiation. The methylation of nine genes (GHSR, ITGA4, RXRG, UTF1, CDH8, FAN19A4, CTNNA2, NEFH, and CASR) was quantitatively validated in 70 pharyngeal SCC cases by pyrosequencing. Hypermethylation significantly correlated with HPV-L1 positivity, but not with age or smoking status. p16INK4A was generally activated in HPV-L1(þ) tumors, and p16-positive cases significantly associated with better prognosis. RXRG hypermethylation strongly correlated with positivity of HPV-L1 and p16 (P ¼ 3 105 and P ¼ 5 104, respectively). RXRG-methylation(þ) significantly associated with better prognosis when analyzing all tumor cases (P ¼ 0.04), and when analyzing the p16-negative poorer-outcome group (P ¼ 0.03). Thus, aberrant DNA methylation might be involved in HPV-associated OPSCC; in addition, DNA methylation could serve as a marker to classify subgroups based on outcome. © 2017 Elsevier B.V. All rights reserved.
Keywords: DNA methylation Human papillomavirus (HPV) Oropharyngeal cancer Hypopharyngeal cancer
Introduction Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide; its annual incidence is 600,000 cases, accounting for 3e5% of all cancers [1]. In particular, the incidence of oropharyngeal squamous cell carcinoma (OPSCC) has increased over the last two decades, mainly because the number of human papillomavirus (HPV)-associated OPSCC cases has risen dramatically [2,3]. HPV-associated and HPV-negative OPSCCs are clinically and biologically distinct. The major risk factors for HPVnegative tumors are tobacco and excessive alcohol use. In
Q1
* Corresponding author. Department of Molecular Oncology, Graduate School of Medicine, Chiba University, 1-8-1 Inohana, Chuo-ku, Chiba 260-8670, Japan. E-mail address: [email protected] (A. Kaneda).
contrast, HPV-associated tumors strongly associate with sexual behavior [4]. Compared to patients with HPV-negative tumors, those with HPV-associated tumors tend to be younger, have higher socioeconomic status, and have disease with lower primary stages and high nodal stages [5]. Regarding prognosis, it was shown that HPV(þ) status is a consistent determinant of better prognosis, regardless of treatment modality, with 5-year survival rates among patients with HPV-associated tumors of approximately 75e80% versus 45e50% among patients with HPV-negative tumors [6]. HNSCCs, especially HPV-negative tumors, are thought to initiate and progress through a series of genetic alterations caused by smoking and excessive alcohol intake. Several tumor suppressive genes are indeed dysregulated through genetic alterations, such as TP53 and CDKN2A, which are associated with the highest mutation rates, based on TCGA data for HNSCC [7]. The prevalence of TP53
http://dx.doi.org/10.1016/j.canlet.2017.08.008 0304-3835/© 2017 Elsevier B.V. All rights reserved.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 2/11
2
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
and CDKN2A mutations in HNSCC ranges from 30 to 84% and 4e74%, respectively, according to various studies [8e10]. However, these two genes are not mutated in HPV-associated tumors; instead, TP53 is inactivated by the E6 oncoprotein [11e14]. HPV induces cancer by infecting epithelial basal cells. Initially after infection, HPV can be identified in circular extra chromosomal particles or episomes. Then, the viral genome typically integrates into the host cell genome, resulting in upregulated expression of viral E6 and E7 oncoproteins [10,15]. E6 and E7 proteins bind to and inactivate the tumor suppressors TP53 and RB1 respectively. This enables the host cells to avoid apoptosis and leads to proliferation [13]. In addition, cell cycle pathway feedback results in overstimulation of cell senescence pathways including p16INK4A. We can therefore use p16INK4A expression as a surrogate marker for HPV infection [16]. HPV infection is generally evaluated by detection of HPV viral oncogenes E6 and E7 using quantitative reverse transcriptase-PCR (RT-PCR) on RNA isolated from fresh frozen samples, or also by detection of L1 region of HPV by PCR combined with immunostaining of p16INK4A [16e19]. DNA methylation is one of the key epigenetic mechanisms that regulates gene expression, and aberrant DNA hypermethylation of gene promoter regions is a major inactivating mechanism to silence tumor suppressor genes in most cancers [20]. In particular, virus infection could induce aberrant DNA methylation during carcinogenesis [21,22]. Therefore, differences in methylation profiles between HPV-associated and HPV-negative OPSCCs have been explored in several studies. Whereas a limited number of similar genes have been investigated, e.g. TIMP3 and p16INK4A [23], detailed DNA methylation profiles of OPSCC have not been analyzed. Moreover, the molecular mechanisms responsible for clinical differences between patients with HPV-associated and HPV-negative tumors have not been clarified. In this study, we performed DNA methylation analysis of OPSCC on a genome-wide scale. We identified candidate methylationassociated genes by Infinium 450k methylome analysis of clinical OPSCC and non-cancerous mucosa samples, and RNA-sequencing (RNA-seq) analysis combined with Infinium 450k analysis for cancer cell lines treated with/without 5-aza-20 -deoxycytidine (Aza) and/or trichostatin A (TSA). We validated methylation status using additional clinical samples by pyrosequencing, and compared methylation status with clinicopathological factors, revealing a significant association with critical factors such as HPV status and prognosis. Materials and methods Full information is provided in the Supplementary Materials and Methods. Clinical samples and cell lines Clinical specimens were collected from 70 patients with pharyngeal squamous cell carcinoma (PSCC) who underwent therapy at Chiba University Hospital or Hamamatsu University hospital and written informed consent was obtained. Biopsy specimens were immediately frozen in liquid nitrogen and stored at 80 C. Frozen materials were microscopically examined by two independent pathologists and were dissected to enrich cancer cells when necessary. Fifty-three samples containing >50% cancer cells were used for subsequent molecular analyses. Seventeen samples were extracted from formalin-fixed paraffin-embedded (FFPE) tissues. The UMSCC47 cell line, derived from a primary tumor of the lateral tongue, was kindly provided by Dr. Thomas E. Carey, University of Michigan. The FaDu cell line, derived from primary hypopharyngeal SCC, was obtained from the ATCC. DNA was extracted by using a QIAquick DNA mini-kit (Qiagen, Hilden, Germany). The protocol was approved by the institutional review board at Chiba University and Hamamatsu University. Immunohistochemistry (IHC) IHC for TP53 was conducted using a DO-7 anti-mouse monoclonal antibody (anti-p53 DO-7 Primary Antibody, Roche, Basel, Switzerland) and the BenchMark ULTRA automated staining system (Roche). Samples with diffuse strong positive nuclear staining were considered TP53-IHC(þ) and designated as TP53-mutation(þ),
those with negative nuclear staining were considered TP53-IHC() and designed as TP53-mutation(þ), and those with sporadic positive nuclear staining were considered TP53-IHC wild-type and designed as TP53-mutation(). For p16INK4A, IHC was performed using an anti-mouse monoclonal antibody (CINtec p16 Histology, Roche) and BenchMark ULTRA, and samples with positive nuclear staining (nuclear and cytoplasmic immunoreactivity in >70% of tumor cells) were considered p16-IHC(þ) and designated as p16(þ). Amplification of L1 DNA region in high risk HPV Evaluation of HPV infection was also performed by amplifying a portion of the L1 region in high risk HPV using GP5 and GP6 primers, as previously reported [24]. Samples with the PCR amplicon was designated as HPV-L1(þ). Aza and TSA treatment UMSCC47 and FaDu cells were seeded at a density of 3 105 cells/10-cm dish on day 0. Cells were exposed to Aza (3 mM for UMSCC47 and 5 mM for FaDu) or vehicle (DMSO) on days 1, 2, and 3. Cells were exposed to TSA (300 nM for UMSCC47 and 100 nM for FaDu) or vehicle (ethanol) on day 3. Medium was changed every 24 h, and cells were harvested on day 4. Infinium assays The Infinium HumanMethylation450 BeadChip (Illumina) contains approximately 485,000 individual CpG sites covering 99% of RefSeq genes with an average of 17 CpG sites per gene. For each CpG site, b-value, ranging from 0.00 to 1.00, was measured by a methylated probe relative to the sum of both methylated and unmethylated probes. Bisulfite conversion was performed using the Zymo EZ DNA Methylation Kit (Zymo Research, Irvine, CA) with 500 ng of genomic DNA for each sample. Whole genome amplification, labeling, hybridization, and scanning were performed according to the manufacturer's protocols. Infinium analysis was performed against 13 clinical OPSCC, four non-cancerous mucosa samples, and cancer cell lines treated with/without Aza and/or TSA (Fig. 1A), and Infinium data were submitted to the GEO database under the accession numbers GSM2612419eGSM2612439. For analysis of DNA methylation at promoter regions based on Infinium data, the probe nearest to the transcription start site (TSS) was selected when multiple probes were designed for one promoter region. The CpG score for each probe was calculated based on a previous report [25,26]. For analysis of CpG island shore regions upstream and downstream from TSS, the probe nearest to TSS was selected among probes at CpG island shore (annotated by Illumina). For analysis of gene body, the probe showing the highest CpG score was selected among probes at gene body, i.e. þ4 kb downstream from TSS through the transcription terminal site. RNA-seq analysis Libraries for RNA-seq were prepared using the TruSeq Stranded mRNA Sample Prep Kit (Illumina, San Diego, CA), following the manufacturer's protocol. Deep sequencing was performed on the Illumina HiSeq 1500 platform using the TruSeq Rapid SBS Kit (Illumina) in 50-base single-end mode according to the manufacturer's protocol. RNA-seq analysis was performed against cancer cell lines treated with/ without Aza and/or TSA (Fig. 1A), and RNA-seq data were submitted to the GEO database under the accession numbers GSM2612408 e GSM2612411. Gene Ontology (GO) analysis Gene annotation enrichment analysis was conducted based on GO (biologic process, cellular component, and molecular function) using the Functional Annotation tool at DAVID Bioinformatics Resources (http://david.abcc.ncifcrf.gov/). Pyrosequencing analysis Quantitative validation of methylation in 70 cases was performed by pyrosequencing using the PyroMark Q96 (Qiagen) as previously reported [26]. Primers were designed by Pyro Q-CpG Software (Qiagen) to amplify bisulfite-treated DNA (Supplementary Table S1). Methylation control samples (0%, 25%, 50%, 75%, and 100%) were prepared as described previously [27] and used to confirm the high quantification ability of pyrosequencing assays. Statistical analysis Association between clinicopathological factors and HPV-L1 status or DNA methylation was analyzed using the Fisher's exact test, c2 test, Student's t-test, ANOVA, and Kolmogorov-Smirnov test. Unsupervised two-way hierarchical clustering was performed using Cluster 3.0 software. Overall Survival (OS) was measured from registration date until date of PSCC-related death, filtering patients alive at last follow-up and non-PSCC-related deaths. Estimation of OS distribution was performed by the Kaplan-Meier method. Differences between groups were determined by a log-rank test using R software (www.r-project.org/). P < 0.05 was considered statistically significant.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 3/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
3
Fig. 1. The design of the study, immunostaining, and HPV status in OPSCC. (A) A flowchart showing the design of this study. (B) Immunostaining of p16 and TP53. Sections of OPSCC tissues were stained with hematoxylin and eosin (left), or analyzed by immunostaining for p16 (middle) and TP53 (right). TP53 was wild type in most p16(þ) OPSCC samples (Case #1). Most of p16() OPSCC cases were TP53-mutation(þ), showing either positive staining of TP53 (Case #2) or negative staining of TP53 (Case #3). Scale bar, 100 mm. (C) HPV infection status. The presence of a 142-bp PCR product for HPV-L1 region was analyzed by gel electrophoresis. Most p16(þ) cases were HPV-L1(þ) (Case #1), whereas p16() cases were generally HPV-L1().
Results Immunostaining analyses and HPV status Expression of p16INK4A and TP53 in HPV-associated OPSCC, HPVnegative OPSCC, and HPSCC was analyzed by IHC (Fig. 1B). HPV
status was evaluated by PCR amplification of L1 region in high risk HPV with GP5/GP6 primers (Fig. 1C). All 22 HPV-associated OPSCC samples that underwent IHC were p16(þ), whereas only one HPVnegative OPSCC specimen was p16(þ). Most p16(þ) cases, except two, were TP53 wild-type.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 4/11
4
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
DNA methylation analysis using Infinium 450k and RNA-seq analysis We performed Infinium 450k BeadArray analysis on 13 clinical OPSCC and four non-cancerous mucosa samples (Fig. 2A). To analyze Infinium data, the probe nearest the TSS was selected for each gene, and b-values were compared by a t-test. A total of 2093 probes in high-CpG promoters had a P-value < 0.05, and these were used for hierarchical clustering. Unsupervised 2-way hierarchical clustering analysis of 13 OPSCC and four normal mucosa samples classified these samples into two clusters. The low methylation cluster included the lowest methylation subgroup, consisting of the four non-cancerous mucosa samples, and a relatively low methylation subgroup of three OPSCC samples. The high methylation cluster included 10 OPSCC cases divided into two subgroups. Interestingly, all five HPV-L1(þ) OPSCC samples were included in the high methylation cluster, forming a unique subcluster, whereas OPSCC samples with relatively low methylation were all HPV-L1(). When a cluster of 487 genes were extracted from the 2093 high-CpG genes, OPSCC samples showed markedly higher methylation levels than non-cancerous samples (0.34 ± 0.00 vs 0.16 ± 0.00, P < 1 1030, t-test) (Fig. 2A). We also investigated CpG island shore regions upstream and downstream from TSS, and gene body regions. The upstream CpG island shore regions were significantly hypermethylated in cancer samples than non-cancerous samples (0.43 ± 0.00 vs 0.35 ± 0.01, P ¼ 3 1030). These regions were already hypermethylated to some extent in non-cancerous mucosa, and the methylation level did not increase in cancer so markedly as observed in high-CpG promoter regions. The downstream CpG island shore was generally hypermethylated in non-cancerous samples, and methylation levels in cancer were not significantly different from non-cancerous samples (0.54 ± 0.00 vs 0.55 ± 0.00, P ¼ 0.2). Gene body showed aberrant hypomethylation in cancer; methylation levels were decreased from 0.72 ± 0.00 in non-cancerous mucosa to 0.61 ± 0.00 in cancer (P < 1 1030) (Fig. 2A). Since methylation levels altered most apparently in high-CpG promoter regions, we focused on these regions for the subsequent analyses. We next extracted frequently hypermethylated and silenced genes in OPSCC samples. Among the 13,221 high-CpG genes with an average b-value <0.2 in the four non-cancerous mucosa samples, 625 were hypermethylated (b-value >0.3) frequently in 3 of 13 OPSCC samples (Fig. 2B). We next performed RNA-seq analysis on UMSCC47 and FaDu cell lines before and after Aza and/or TSA treatment, and found 3595 and 2778 genes showing >2-fold upregulation in UMSCC47 and FaDu cell lines, respectively, compared to expression levels in untreated cells. Among the 625 frequently hypermethylated genes, 144 were highly methylated (bvalue >0.6) in at least one cell line before Aza/TSA treatment, with a >2-fold upregulation after Aza/TSA treatment, accompanying a decrease in the b-value (Fig. 2B). For these 144 genes, we confirmed the unmethylated status in all four non-cancerous mucosa samples and the hypermethylated status in OPSCC cases in the surrounding probes as well as the selected probe nearest to the TSS (Fig. 2C). GO terms related to cellcell signaling and neuron differentiation were significantly enriched (Fig. 2D). Validation of DNA methylation by pyrosequencing Nine genes were chosen from the 144 frequently silenced genes, including those related to relevant GO terms, and methylation levels were validated by quantitative methylation analysis by pyrosequencing. First, we analyzed the methylation level at the
selected probe site using OPSCC and non-cancerous mucosa samples that were previously analyzed by Infinium (Fig. 3A). Methylation levels were confirmed to be highly similar. p16INK4A was additionally analyzed, and confirmed to be unmethylated in both non-cancerous mucosa and OPSCC samples. We next investigated the methylation status of these nine genes in pharyngeal cancer samples including 42 OPSCC and 15 hypo-pharyngeal SCC by pyrosequencing (Fig. 3B). Not only the selected probe site, but also the surrounding CpG sites were similarly hypermethylated, indicating that aberrant DNA methylation does not occur randomly at a single CpG site, but occurs at multiple consecutive CpG sites in a region near the TSS. When tumor samples were ordered (from the left) based on average methylation level, HPV-L1(þ) samples were significantly enriched on the left (P ¼ 0.01, Kolmogorov-Smirnov test). This indicated that HPV-positivity significantly correlates with frequent DNA hypermethylation. However, other clinicopathological factors such as age or smoking status were not significantly correlated with DNA methylation level. When DNA methylation levels were compared between HPV-L1(þ) and HPVL1() samples, significantly higher methylation was detected in all the cancer samples (P ¼ 0.0002, t-test) or in OPSCC samples (P ¼ 0.001, t-test). Comparison of HPV status, IHC data, and methylation status As p16INK4A is a surrogate marker of HPV infection, p16(þ) status correlated strongly with HPV-L1(þ). TP53-mutation(þ) cases included 21 cases with nuclear staining and 16 cases with negative staining, and these were strongly, inversely correlated with p16(þ) cases (Fig. 4A). To compare methylation status with HPV status and IHC data, we investigated the methylation status of the four most frequently methylated genes (RXRG, GHSR, CTNNA2, and ITGA4). Aberrant methylation was significantly correlated with HPV-L1(þ) cases, and RXRG methylation showed a particularly strong correlation with HPV-positivity (P ¼ 3 105, c2-test) (Fig. 4B). Despite the significant correlation, not only p16(þ) cases, but also some p16() cases showed aberrant RXRG methylation. We also analyzed the association between clinicopathological factors and HPV-L1 status, or RXRG methylation levels (Table 1). HPV-L1 positivity associated significantly with tumor location (tonsils), p16(þ), wild-type TP53, and alcohol() (P ¼ 0.002, 2 1012, 3 108, 0.002, respectively). These associations were in good agreement with the previous reports that HPV-associated OPSCC arises within tonsils and base of tongue, that HPVassociated OPSCC patients do not intake excessive alcohol, and that TP53 mutation is generally found in HPV-negative cases, not in HPV-associated cases [28]. High RXRG methylation levels also associated significantly with tumor location (tonsils), p16(þ), and wild-type TP53 (P ¼ 0.04, 5 104, 0.002, respectively). We then investigated the association with prognosis. p16(þ) cases and L1(þ) cases, which are suggested to be HPV-associated tumors, significantly associated with better prognosis than p16() cases (P ¼ 0.002, log-rank test) and L1() cases (P ¼ 0.005), respectively (Fig. 5A). Whereas TP53-mutation did not predict outcome, RXRG-methylation(þ) cases significantly associated with better prognosis when analyzing all cancer cases (P ¼ 0.04) (Fig. 5A). When analyzing a p16(þ) better-outcome group and a p16() poorer-outcome group separately, all the p16(þ) cases showed good prognosis regardless of RXRG methylation status. Among p16() poorer-outcome patients, RXRG-methylation(þ) correlated significantly with good prognosis (P ¼ 0.03) (Fig. 5B). From these results, two-step panel might be generated to predict prognosis. The first panel “p16 IHC” extracts p16(þ) better-outcome patients, and the second panel “RXRG methylation” additionally
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 5/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
5
Fig. 2. Infinium 450k analysis of OPSCC and non-cancerous samples. (A) Genome-wide DNA methylation analyzed for 13 clinical OPSCC and four non-cancerous samples. A probe nearest to the transcriptional start site (TSS) was selected for analysis of each promoter region. Probes with significantly different b-values between OPSCC and non-cancerous samples (P < 0.05, t-test) were extracted, and 2093 genes with high-CpG promoters (CpG score >0.48) were used for clustering (top). Genes with low-CpG promoters (CpG score <0.48), CpG island shore regions upstream and downstream from TSS, and gene body regions are additionally shown (middle). White: HPV-L1(), age<60, smoking(); black: HPV-L1(þ), age>60, smoking(þ) (bottom). Whereas three OPSCC samples with relatively low methylation formed a cluster with non-cancerous samples, other OPSCC samples comprised a cluster with high methylation. Methylation levels between tumor and non-cancerous samples were compared using t-test (right panels). The difference of methylation levels was the most apparent in high-CpG promoter region (P < 1 1030, t-test). (B) There were 144 genes that met the following criteria: (i) unmethylated in four non-cancerous samples (average bvalue <0.2); (ii) frequently hypermethylated in 3 of 13 clinical OPSCC samples (b-value >0.3); (iii) hypermethylated in at least one cancer cell line before Aza/TSA treatment (b-value >0.6); (iv) de-methylated after Aza/TSA treatment (i.e. showing a decreased of b-value). (C) For the 144 frequently silenced genes, we confirmed that they were unmethylated in the four non-cancerous mucosa samples and hypermethylated in OPSCC cases in the surrounding probes as well as the selected probe nearest to the TSS. Representatively, GHSR was unmethylated in four non-cancerous samples (N), but OPSCC samples (T) were hypermethylated at the probe site nearest to the TSS and the surrounding probe sites as well. Arrowhead, the position of TSS. (D) For these genes, Gene Ontology (GO) terms related to cell-cell signaling and neuron differentiation were significantly enriched.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 6/11
6
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
Fig. 3. Validation of DNA methylation in OPSCC samples by pyrosequencing. (A) Nine genes were chosen from the 144 frequently silenced genes. The methylation levels (analyzed by Infinium) are shown in the color scale (top), and the methylation levels at the probe site was also analyzed by pyrosequencing and shown (bottom), confirming similar results. p16 was additionally analyzed. (B) Pyrosequencing data of 70 pharyngeal cancer samples. To investigate the methylation status in other pharyngeal cancer samples and the relationship with clinicopathological factors (bottom), methylation levels in 42 OPSCC and 15 hypo-pharyngeal SCC samples were additionally analyzed by pyrosequencing (top). Closed boxes: HPV-L1(þ), oropharyngeal, elder age >60 years old, smoking(þ), respectively. Tumor samples were ordered from left to right based on average methylation level of the nine genes; HPV-L1(þ) samples were significantly enriched on the left (P ¼ 0.01, Kolmogorov-Smirnov test), i.e. HPV infection correlated significantly with frequent hypermethylation. HPVL1(þ) samples were associated with higher methylation levels within all tumor samples (P ¼ 0.0002, t-test) and also within OPSCC (P ¼ 0.001, t-test). Frequent hypermethylation did not correlate with age or smoking status. Whereas red numbers indicated the CpG sites of the analyzed Infinium probes, the surrounding CpG sites as well as the probe site were shown to be hypermethylated at similar levels. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 7/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
7
Fig. 4. Comparison of HPV status, immunostaining data, and methylation status in OPSCC samples. (A) Left, HPV-L1 status and IHC data of p16 and TP53 (See Fig. 1 for representative results). Right, the methylation status of the four most frequently methylated genes (RXRG, GHSR, CTNNA2, and ITGA4). White, methylation level <50%. Black, methylation level 50%. HPV-L1(þ) status correlated strongly with p16(þ) status (P < 1 1015, Fisher's exact test), and TP53-mutation(þ) cases showed a strong inverse correlation with p16(þ) cases (P ¼ 2 1010). (B) Correlation between aberrant methylation and HPV-L1(þ) status. For each of the four genes, especially RXRG, aberrant methylation significantly correlated with HPV-L1(þ) samples.
extracts RXRG-methylation(þ) better-outcome patients among p16() group (Fig. 5C). Discussion The involvement of epigenetic alterations in OPSCC has not been clarified; thus, we analyzed aberrant DNA methylation in OPSCC on
a genome-wide scale using the Infinium 450k BeadArray. We identified frequently hypermethylated and silenced genes in OPSCC, such as RXRG, and showed that these genes are preferentially hypermethylated in HPV-positive tumors. p16INK4A is activated in HPV-positive tumors, and p16(þ) cases show significantly better prognosis than p16() cases. Moreover, RXRG-methylation(þ) cases significantly associated with better prognosis when
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 8/11
8
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
Table 1 Clinicopathological factors and their association with DNA methylation. All cases
# of cases 70 Gender Male 60 Female 10 Age >60 51 <60 19 Tumor location tonsils 34 base of tongue 8 soft palate 11 posterior wall 2 Clinical tumor classification T1 6 T2 33 T3 16 T4a 14 T4b 1 Clinical lymph node classification N0 23 N1 4 N2a 3 N2b 25 N2c 12 N3 3 Lymph node metastasis (þ) 47 () 23 Clinical stage I 6 II 11 III 8 IVA 39 IV B 6 p16 IHC (þ) 23 () 39 N/A 8 TP53 IHC wild type 25 mutation(þ) 37 N/A 8 Smoking status (þ) 51 () 9 N/A 10 Alcohol status (þ) 47 () 12 N/A 11
HPSCC
OPSCC
RXRG methylation level
L1(þ)
L1()
P-value
Methylation (%)
P-value
15
24
31
12 3
20 4
28 3
0.7
36.3 ± 3.1 46.9 ± 10.2
0.4
9 6
18 6
24 7
1
38.4 ± 3.6 37.5 ± 6.0
0.9
e e e e
21 3 0 0
13 5 11 2
0.002*
47.3 35.8 30.0 44.4
± ± ± ±
4.1 9.5 4.2 41.6
0 3 5 7 0
0 16 5 2 1
6 14 6 5 0
± ± ± ±
8.2 4.8 6.3 6.6
0.2
32.9 43.5 32.6 34.2 55
3 2 0 6 4 0
6 1 3 9 3 2
14 1 0 10 5 1
0.3
30.4 30.2 62.8 40.0 41.0 57.8
± ± ± ± ± ±
12 3
18 6
17 14
0.2
41.8 ± 3.6 30.4 ± 5.5
0 0 4 11 0
0 5 2 12 5
6 6 2 16 1
0.07
32.9 30.6 30.3 39.9 58.0
0 14 1
23 0 1
1 24 6
2 10e12*
52.9 ± 5.2 29.4 ± 3.5
5 10e4*
4 10 1
20 2 2
3 23 5
3 10e8*
50.7 ± 5.3 29.6 ± 3.5
0.002*
8 1 6
17 6 1
26 2 3
0.1
41.3 ± 3.4 43.4 ± 10.4
0.9
9 0 6
12 10 2
26 2 3
0.002*
39.8 ± 3.7 50.7 ± 7.0
0.2
0.04*
0.7
± ± ± ± ±
5.5 13.7 13.9 4.6 8.3 5.0
8.2 10.5 8.8 4.1 2.0
0.09
0.1
0.09
OPSCC, oropharyngeal squamous cell carcinoma. HPSCC, hypo-pharyngeal squamous cell carcinoma. N/A, not applicable. For OPSCC, association with HPV-L1 status was analyzed by Fisher's exact test or c2 test. For RXRG methylation, methylation levels were compared by t-test or ANOVA. *P < 0.05.
analyzing all tumor cases, and also when analyzing the p16() poorer-outcome group. More than 10 genes were previously reported to be inactivated by DNA methylation in OPSCC [11,29,30], including p16INK4A, HOXA9, MGMT, GATA4, and RASSF1 [23,30]. Herein, 144 genes were extracted as frequently hypermethylated and silenced in OPSCC, but p16INK4A was rarely methylated and was not included. GO terms related to neuron differentiation and cell development were enriched in these genes, in agreement with previous studies reporting that genes related to differentiation and development are targets of the polycomb complex in ES cells and that these genes are preferentially hypermethylated in many types of cancer [26,31]. Lechner et al. reported 43 hypermethylated promoter regions associated with HPV in HNSCC through Infinium analysis, including three cadherins of the Polycomb group target genes [32]. Weiss
et al. investigated promoter methylation of 12 genes (TIMP3, CDH1, p16INK4A, DAPK1, TCF21, CD44, MLH1, MGMT, RASSF1, CCNA1, LARS2, and CEPBA) by methylation-specific PCR, which is not a quantitative method, in a cohort of 55 HNSCC cases derived mostly from the oropharynx. HPV-associated tumors significantly correlated with CCNA1 and TIMP3 promoter methylation [23]. Schlecht et al. reported 22 CpG loci showing significant difference of methylation level between HPV-positive and HPV-negative OPSCC, which included four CpG loci at CDKN2A downstream region, but not promoter region around TSS [33]. O'Regan et al. showed frequent p16INK4A hypermethylation in HNSCC of anterior two thirds of the tongue, but no p16INK4A methylation in OPSCC [34]. The association of HPV-positive OPSCC with DNA hypermethylation of promoter regions, including p16 promoters, is thus still controversial. In this study, a highly quantitative analysis based on pyrosequencing
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 9/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
9
Fig. 5. Analysis of prognosis in pharyngeal cancer cases. (A) Analysis of prognosis based on p16, HPV-L1, TP53 mutation, and RXRG methylation status. p16(þ) cases were significantly associated with better prognosis than p16() cases as reported (P ¼ 0.002, log-rank test), and HPV-L1(þ) cases that were mostly p16(þ) were also significantly correlated with better outcome (P ¼ 0.005). Whereas TP53-mutation did not predict outcome, RXRG-methylation(þ) cases were significantly associated with better prognosis when analyzing all cancer cases (P ¼ 0.04). (B) RXRG methylation correlated with better outcome even when analyzing the p16() poorer-outcome group (P ¼ 0.03). p16(þ) cases showed good prognosis regardless of RXRG methylation status. (C) Two-step panel to predict the prognosis of OPSCC. The first and second panels classify patients into a good prognosis group with p16(þ) and/or RXRG-methylation(þ), and a poor prognosis group with p16() and RXRG-methylation().
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
CAN13454_proof ■ 19 August 2017 ■ 10/11
10
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
revealed that hypermethylation of p16INK4A promoter region is very rare. Instead, promoter regions of 144 genes were found to be frequently hypermethylated in OPSCC, and analysis of nine genes showed that frequent DNA methylation is significantly correlated with HPV-positive status. Other than promoter CpG islands, CpG island shore and gene body regions were also analyzed. Gene body generally showed high methylation levels in non-cancerous mucosa, and significant decrease of methylation levels was observed in OPSCC. This is consistent with global DNA hypomethylation reported in many types of cancer [35]. CpG island shore regions upstream from TSS were hypermethylated in OPSCC, but the difference of methylation levels between cancer and non-cancerous samples was smaller than promoter regions. Aberrant DNA methylation is reported to occur from lower-CpG peripheral regions to higher-CpG regions in high-CpG promoters in carcinogenesis, and aberrant methylation at CpG island shore is frequently observed in the surrounding noncancerous tissues [36,37]. The expression of p16INK4A is used as a surrogate marker of HPV infection. E6 and E7 oncoproteins are known to inactivate TP53 and RB1, which enables the host cells to avoid apoptosis and proliferate; this simultaneously overstimulates cellular senescence pathways because of cell cycle pathway feedback. p16INK4A is thus strongly expressed in HPV-positive OPSCC. Our data also indicated that hypermethylation of p16INK4A is not the cause of p16() status in HPV-negative cases. It has been reported that HPV-negative HNSCC featured novel focal deletions in CDKN2A [28]. Further analysis is necessary to clarify whether negative expression of p16 is due to gene mutation or other causes. Interestingly, although there was a significant correlation between DNA hypermethylation and HPVpositivity, a fraction of HPV-negative OPSCC also exhibited DNA hypermethylation. Whereas p16 positivity (thus HPV positivity) is a marker of better OPSCC prognosis [5,6], RXRG-methylation(þ) cases showed better outcome among all tumor cases, and even among the p16() poorer-outcome group. Retinoid X receptors (RXRs) are nuclear receptors for retinoids that have important roles in the regulation of growth and differentiation. RXRs have three separate isotypes encoded on three separate genes: RXRA, RXRB, and RXRG. Promoter hypermethylation of RXRG and the loss of gene expression were reported in lung cancer [38]. Although whether RXRG methylation plays a critical role in carcinogenesis is yet to be investigated, we identified frequent hypermethylation of RXRG in OPSCC and its significant correlation with better prognosis. Further studies are necessary to clarify how HPV infection is associated with DNA hypermethylation, or alternatively, whether there is another independent mechanism that confers relatively better outcome in OPSCC with higher DNA methylation. It is also reported that exposure to selective RXRG ligands decrease cellular proliferation of thyroid cancer cell lines expressing RXRG [39,40]. Although further studies should be investigated, RXRG might be a potential target for treatment of RXRG-expressing cancer [41]. In summary, in OPSCC, we identified frequently hypermethylated and silenced genes that are preferentially hypermethylated in HPV-associated tumors. While p16INK4A expression is known to be a marker of better prognosis in OPSCC, RXRG methylation(þ) cases show better prognosis among all the cases as well as within p16() cases. Acknowledgements We thank Chihomi Sato, Keiko Iida, and Aki Komatsu for technical assistance, and Editage (www.editage.jp) for English language editing. This study was supported by Japan Agency for Medical Research and Development (AMED) [Practical Research for
Innovative Cancer Control 17ck0106263h0001 to AK]; Japan Society for the Promotion of Science (JSPS) [24249081 to YO, 16H05412 to AK]; and a grant from Global and Prominent Research, Chiba University to AK. Q2 Conflict of interest All authors have no conflicts of interest to disclose. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.canlet.2017.08.008. References [1] J. Ferlay, H.R. Shin, F. Bray, D. Forman, C. Mathers, D.M. Parkin, Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008, Int. J. Cancer 127 (2010) 2893e2917. [2] A.K. Chaturvedi, E.A. Engels, W.F. Anderson, M.L. Gillison, Incidence trends for human papillomavirus-related and -unrelated oral squamous cell carcinomas in the United States, J. Clin. Oncol. 26 (2008) 612e619. [3] A.B. Ryerson, E.S. Peters, S.S. Coughlin, V.W. Chen, M.L. Gillison, M.E. Reichman, et al., Burden of potentially human papillomavirus-associated cancers of the oropharynx and oral cavity in the US, 1998-2003, Cancer 113 (2008) 2901e2909. [4] M.L. Gillison, G. D'Souza, W. Westra, E. Sugar, W. Xiao, S. Begum, et al., Distinct risk factor profiles for human papillomavirus type 16-positive and human papillomavirus type 16-negative head and neck cancers, J. Natl. Cancer Inst. 100 (2008) 407e420. [5] K. Grisar, R. Dok, J. Schoenaers, T. Dormaar, E. Hauben, M. Jorissen, et al., Differences in human papillomavirus-positive and -negative head and neck cancers in Belgium: an 8-year retrospective, comparative study, Oral Surg. Oral Med. Oral Pathol. Oral Radiol. 121 (2016) 456e460. [6] K.K. Ang, J. Harris, R. Wheeler, R. Weber, D.I. Rosenthal, P.F. Nguyen-Tan, et al., Human papillomavirus and survival of patients with oropharyngeal cancer, N. Engl. J. Med. 363 (2010) 24e35. [7] N. Agrawal, M.J. Frederick, C.R. Pickering, C. Bettegowda, K. Chang, R.J. Li, et al., Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1, Science 333 (2011) 1154e1157. [8] H. Blons, P. Laurent-Puig, TP53 and head and neck neoplasms, Hum. Mutat. 21 (2003) 252e257. [9] D.N. Hayes, C. Van Waes, T.Y. Seiwert, Genetic landscape of human papillomavirus-associated head and neck cancer and comparison to tobaccorelated tumors, J. Clin. Oncol. 33 (2015) 3227e3234. [10] W. Zhang, A. Edwards, Z. Fang, E.K. Flemington, K. Zhang, Integrative genomics and transcriptomics analysis reveals potential mechanisms for favorable prognosis of patients with HPV-positive head and neck carcinomas, Sci. Rep. 6 (2016) 24927. [11] P.M. van Kempen, R. Noorlag, W.W. Braunius, I. Stegeman, S.M. Willems, W. Grolman, Differences in methylation profiles between HPV-positive and HPV-negative oropharynx squamous cell carcinoma: a systematic review, Epigenetics 9 (2014) 194e203. [12] N. Stransky, A.M. Egloff, A.D. Tward, A.D. Kostic, K. Cibulskis, A. Sivachenko, et al., The mutational landscape of head and neck squamous cell carcinoma, Science 333 (2011) 1157e1160. [13] C.H. Chung, M.L. Gillison, Human papillomavirus in head and neck cancer: its role in pathogenesis and clinical implications, Clin. Cancer Res. 15 (2009) 6758e6762. [14] M. Lechner, G.M. Frampton, T. Fenton, A. Feber, G. Palmer, A. Jay, et al., Targeted next-generation sequencing of head and neck squamous cell carcinoma identifies novel genetic alterations in HPVþ and HPV tumors, Genome Med. 5 (2013) 49. [15] M. Parfenov, C.S. Pedamallu, N. Gehlenborg, S.S. Freeman, L. Danilova, C.A. Bristow, et al., Characterization of HPV and host genome interactions in primary head and neck cancers, Proc. Natl. Acad. Sci. U. S. A. 111 (2014) 15544e15549. [16] C. Ndiaye, M. Mena, L. Alemany, M. Arbyn, X. Castellsague, L. Laporte, et al., HPV DNA, E6/E7 mRNA, and p16INK4a detection in head and neck cancers: a systematic review and meta-analysis, Lancet Oncol. 15 (2014) 1319e1331. [17] M.M. Rietbergen, C.R. Leemans, E. Bloemena, D.A. Heideman, B.J. Braakhuis, A.T. Hesselink, et al., Increasing prevalence rates of HPV attributable oropharyngeal squamous cell carcinomas in the Netherlands as assessed by a validated test algorithm, Int. J. Cancer 132 (2013) 1565e1571. [18] M.A. Broglie, W. Jochum, A. Michel, T. Waterboer, D. Foerbs, R. Schoenegg, et al., Evaluation of type-specific antibodies to high risk-human papillomavirus (HPV) proteins in patients with oropharyngeal cancer, Oral Oncol. 70 (2017) 43e50. [19] S.J. Smeets, A.T. Hesselink, E.J. Speel, A. Haesevoets, P.J. Snijders, M. Pawlita, et al., A novel algorithm for reliable detection of human papillomavirus in
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
Q3
66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
CAN13454_proof ■ 19 August 2017 ■ 11/11
T. Nakagawa et al. / Cancer Letters xxx (2017) 1e11
[20] [21]
[22]
[23]
[24]
[25]
[26]
[27]
[28] [29]
[30]
paraffin embedded head and neck cancer specimen, Int. J. Cancer 121 (2007) 2465e2472. D.J. Weisenberger, Characterizing DNA methylation alterations from the cancer genome atlas, J. Clin. Invest. 124 (2014) 17e23. A. Kaneda, K. Matsusaka, H. Aburatani, M. Fukayama, Epstein-Barr virus infection as an epigenetic driver of tumorigenesis, Cancer Res. 72 (2012) 3445e3450. Z. Herceg, A. Paliwal, Epigenetic mechanisms in hepatocellular carcinoma: how environmental factors influence the epigenome, Mutat. Res. 727 (2011) 55e61. D. Weiss, T. Basel, F. Sachse, A. Braeuninger, C. Rudack, Promoter methylation of cyclin A1 is associated with human papillomavirus 16 induced head and neck squamous cell carcinoma independently of p53 mutation, Mol. Carcinog. 50 (2011) 680e688. K. Husnjak, M. Grce, L. Magdic, K. Pavelic, Comparison of five different polymerase chain reaction methods for detection of human papillomavirus in cervical cell specimens, J. Virol. Methods 88 (2000) 125e134. M. Weber, I. Hellmann, M.B. Stadler, L. Ramos, S. Paabo, M. Rebhan, et al., Distribution, silencing potential and evolutionary impact of promoter DNA methylation in the human genome, Nat. Genet. 39 (2007) 457e466. K. Matsusaka, A. Kaneda, G. Nagae, T. Ushiku, Y. Kikuchi, R. Hino, et al., Classification of Epstein-Barr virus-positive gastric cancers by definition of DNA methylation epigenotypes, Cancer Res. 71 (2011) 7187e7197. K. Yagi, K. Akagi, H. Hayashi, G. Nagae, S. Tsuji, T. Isagawa, et al., Three DNA methylation epigenotypes in human colorectal cancer, Clin. Cancer Res. 16 (2010) 21e33. N. Cancer Genome Atlas, Comprehensive genomic characterization of head and neck squamous cell carcinomas, Nature 517 (2015) 576e582. R. Guerrero-Preston, C. Michailidi, L. Marchionni, C.R. Pickering, M.J. Frederick, J.N. Myers, et al., Key tumor suppressor genes inactivated by “greater promoter” methylation and somatic mutations in head and neck cancer, Epigenetics 9 (2014) 1031e1046. N.V. Anayannis, N.F. Schlecht, T.J. Belbin, Epigenetic mechanisms of human papillomavirus-associated head and neck cancer, Arch. Pathol. Lab. Med. 139 (2015) 1373e1378.
11
[31] M. Widschwendter, H. Fiegl, D. Egle, E. Mueller-Holzner, G. Spizzo, C. Marth, et al., Epigenetic stem cell signature in cancer, Nat. Genet. 39 (2007) 157e158. [32] M. Lechner, T. Fenton, J. West, G. Wilson, A. Feber, S. Henderson, et al., Identification and functional validation of HPV-mediated hypermethylation in head and neck squamous cell carcinoma, Genome Med. 5 (2013) 15. [33] N.F. Schlecht, M. Ben-Dayan, N. Anayannis, R.A. Lleras, C. Thomas, Y. Wang, et al., Epigenetic changes in the CDKN2A locus are associated with differential expression of P16INK4A and P14ARF in HPV-positive oropharyngeal squamous cell carcinoma, Cancer Med. 4 (2015) 342e353. [34] E.M. O'Regan, M.E. Toner, S.P. Finn, C.Y. Fan, M. Ring, B. Hagmar, et al., p16(INK4A) genetic and epigenetic profiles differ in relation to age and site in head and neck squamous cell carcinomas, Hum. Pathol. 39 (2008) 452e458. [35] M. Ehrlich, DNA hypomethylation in cancer cells, Epigenomics 1 (2009) 239e259. [36] K. Matsusaka, S. Funata, M. Fukuyo, Y. Seto, H. Aburatani, M. Fukayama, et al., Epstein-Barr virus infection induces genome-wide de novo DNA methylation in non-neoplastic gastric epithelial cells, J. Pathol. (2017). Q4 [37] R.A. Irizarry, C. Ladd-Acosta, B. Wen, Z. Wu, C. Montano, P. Onyango, et al., The human colon cancer methylome shows similar hypo- and hypermethylation at conserved tissue-specific CpG island shores, Nat. Genet. 41 (2009) 178e186. [38] S.M. Lee, J.Y. Lee, J.E. Choi, S.Y. Lee, J.Y. Park, D.S. Kim, Epigenetic inactivation of retinoid X receptor genes in non-small cell lung cancer and the relationship with clinicopathologic features, Cancer Genet. Cytogenet 197 (2010) 39e45. [39] J. Klopper, M. Kane, A. Jimeno, S. Sams, J. French, L. Pike, et al., A phase II trial of bexarotene for advanced differentiated thyroid cancer, Thyroid 25 (2015) 563e564. [40] J.P. Klopper, W.R. Hays, V. Sharma, M.A. Baumbusch, J.M. Hershman, B.R. Haugen, Retinoid X receptor-gamma and peroxisome proliferatoractivated receptor-gamma expression predicts thyroid carcinoma cell response to retinoid and thiazolidinedione treatment, Mol. Cancer Ther. 3 (2004) 1011e1020. [41] B.R. Haugen, L.L. Larson, U. Pugazhenthi, W.R. Hays, J.P. Klopper, C.A. Kramer, et al., Retinoic acid and retinoid X receptors are differentially expressed in thyroid cancer and thyroid carcinoma cell lines and predict response to treatment with retinoids, J. Clin. Endocrinol. Metab. 89 (2004) 272e280.
Please cite this article in press as: T. Nakagawa, et al., Frequent promoter hypermethylation associated with human papillomavirus infection in pharyngeal cancer, Cancer Letters (2017), http://dx.doi.org/10.1016/j.canlet.2017.08.008
30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58