Validation of nucleolar protein 4 as a novel methylated tumor suppressor gene in head and neck cancer

Methylation of CpG islands in the promoter region of genes acts as a significant mechanism of epigenetic gene silencing in head and neck cancer. In the present study, we assessed the association of epigenetic alterations of a panel of 12 genes [nucleolar protein 4 (NOL4), iroquois homeobox 1 (IRX1), SLC5A8, LRRC3B, FUSSEL18, EBF3, GBX2, HMX2, SEPT9, ALX3, SOCS3 and LHX6] with head and neck squamous cell carcinoma (HNSCC) via a candidate gene approach. After the initial screening of methylated CpG islands on the promoter regions by bisulfite sequencing using salivary rinse samples, only two genes had methylated CpG dinucleotides on their promoter regions in tumor samples and absence of methylated CpGs were found in normal salivary rinse samples after bisulfite modification and bisulfite sequencing. We then performed real-time quantitative methylation-specific PCR (QMSP) on 16 salivary rinse and 14 normal mucosal samples from healthy subjects and 33 HNSCC tumor samples for the two genes selected. After validation with QMSP, one gene, NOL4, was highly methylated (91%) in tumor samples and unmethylated in normal salivary rinses and minimally methylated in normal mucosal samples demonstrating cancer-specific methylation in HNSCC tissues. Although the IRX1 gene was observed as methylated in normal mucosal and salivary rinse samples, the methylation values of these normal samples were very low (<10%). In conclusion, we identified NOL4 as a highly specific promoter methylated gene associated with HNSCC. IRX1 may have potential as a biomarker for HNSCC and should be assessed in a larger cohort.


Introduction
Head and neck cancer, which is the sixth most common cancer in the world among human malignant disorders, is an aggressive and life-threatening disease with poor prognosis, morbidity and high mortality in advanced disease. Survival rates have not improved significantly for patients with head and neck squamous cell carcinoma (HNSCC) in the past 30 years despite active clinical and basic research addressing this issue. More than 40,000 new cases of HNSCC are diagnosed in the United States each year, with a mortality rate of 12,000 in the USA annually (1). Treatment for HNSCC includes surgical resection, chemotherapy and radiation therapy; however, approximately 50% of all patients have advanced disease at the time of diagnosis often requiring use of all three treatment modalities. Cancer-specific molecular biomarkers, which have the ability to warn the clinicians in the earlier stage before the disease advances, or to provide insight regarding the prognosis of the disease or outcome of the patients, are required. In addition, it is important to develop new methods that provide sensitive and reliable biomarkers of HNSCC for detection, treatment response and prognosis.
Epigenetic alterations are a recent attractive phenomenon of human cancer, with the activation of proto-oncogenes and inactivation of tumor suppressor genes, either through hypomethylation or hypermethylation in the promoter regions of the genes, respectively (2). Transcriptional silencing of tumor suppressor genes by means of promoter hypermethylation plays an important role in head and neck carcinogenesis (3). Methylation of the CpG islands in the promoter regions of tumor suppressor genes is frequently observed with reduced gene expression (4,5).

Materials and methods
Tissue samples. Following institutional review board approval and after obtaining appropriate informed consent, the HNSCC patients and control population from healthy subjects enrolled in a community screening study were recruited from the Johns Hopkins School of Medicine, Department of Otolaryngology-Head and Neck Surgery. Mucosal samples and salivary rinses from healthy population and HNSCC tissue samples were collected. In the present study, salivary rinses were obtained by brushing oral cavity and oropharyngeal surfaces with an exfoliating brush followed by rinse and gargle with 20 ml normal saline solution. The brush was gently agitated to release the obtained material into saline. After centrifugation, the supernatant was discarded and DNA was isolated from the pellet. Tumors were snap frozen and microdissected on a cryostat to ≥75% purity. DNA from 16 salivary rinse samples from noncancer individuals were analyzed as a control, to investigate the normal promoter methylation status of two newly identified candidate genes, IRX1 and NOL4. The methylation status of these genes was analyzed in 33 fresh tumor samples from patients with head and neck cancer and 14 normal mucosa samples from healthy individuals.
DNA extraction and bisulfite treatment. DNA was isolated as previously described (28). In brief, DNA was obtained by phenol/chloroform extraction after overnight incubation with proteinase K (Boehringer-Mannheim, Germany) at 48˚C. DNA from tumor and control samples was subjected to bisulfite treatment using EpiTect Bisulfite modification kit (Qiagen, Valencia, CA, USA) as per the manufacturer's protocol.
Bisulfite sequencing. The bisulfite sequence analysis was performed to determine the methylation status in the promoter regions of 12 genes. Bisulfite-treated DNA was amplified for the 5' region that included at least a portion of the CpG island within 1-2 kb of the first exon of the genes. The promoter regions of the genes were found from the database of the University of California, Santa Cruz (UCSC) (http:// genome.ucsc.edu/). Primer sequences were determined by MethPrimer program (29) showing the CpG islands in the promoter regions of 12 genes for bisulfite sequencing (Table I).
A thousand base pair region of the genes' promoters and some part of the first exon were sequenced by the specific primers producing 400-500 bp PCR fragments. The primers for bisulfite sequencing were designed to hybridize to regions in the promoter without CpG dinucleotides. PCR products were gel-purified using the QIAquick Gel Extraction kit (Qiagen) according to the manufacturer's instructions. Each amplified DNA sample was sequenced by the Applied Biosystems 3700 DNA Analyzer using nested, forward or reverse primers and BD terminator dye (Applied Biosystems, Foster City, CA, USA).
Quantitative methylation-specific PCR. Primer and probe sequences were determined by MethPrimer program showing the CpG islands in the promoter regions of two genes selected after bisulfite sequencing (Table II). To determine if the methylated genes in tumor samples were cancer-specific, we investigated promoter methylation in 16 normal saliva, 14 age-matched normal mucosa from healthy individuals that were analyzed as a control, to investigate the normal promoter methylation status of two newly identified candidate genes (NOL4, IRX1) and in 33

Results
Clinicopathological characteristics of control subjects and patients with HNSCC. Genes specifically methylated in HNSCC tumors. In the present study, we investigated methylation of the 12 gene promoters by bisulfite modification and QMSP. In the initial screening of the methylated CpG islands, bisulfite sequencing was performed by using four HNSCC cell lines (JHU-06, JHU-022, JHU-022B and JHU-028), eight normal salivary rinses and eight HNSCC samples. Only two genes, NOL4 (NM_003787) and IRX1 (NM_024337) (http://www. genenames.org), had methylated CpG dinucleotides on their promoter regions in tumor samples but absence of methylated CpGs was found in normal salivary rinse samples (Table IV). We investigated the methylation frequency in a larger cohort of normal salivary rinses, mucosal and HNSCC specimens, in order to find a biomarker candidate.
In the second stage of the present study, we performed QMSP on 16 normal salivary rinse and 14 normal mucosal samples from healthy individuals and 33 HNSCC tumor samples for two selected genes. The NOL4 gene showed no methylation (0/16) in normal salivary rinses and two out of 14 (14%) mucosal samples were minimally methylated between 1 and 3% methylation values, whereas the methylation rate was 91% (30/33) on the promoter region of the NOL4 gene in   (Fig. 1). Although normal salivary methylation rate was observed as high, IRX1 methylation values were <1% in normal salivary rinses and were mostly (8/14, 57%) between 0.1 and 10% in normal mucosal samples indicating a very strong marker which may define HNSCC tumor tissues as a potential biomarker (Fig. 1).

Discussion
In the present study, we investigated methylation status of the promoter regions of 12 genes by bisulfite modification, bisulfite sequencing and QMSP techniques. Only two of them (IRX1 and NOL4) showed methylation in their CpG islands located on promoter region of these genes, indicating a characteristic of a biomarker molecule. NOL4 gene encoding a nucleolar protein which is expressed predominantly in brain and testis, was identified by Ueki et al (31) and there is only one study showing high methylation status of 20 patients with cervical cancer by MethyLight assays (32) in concordance with our results.
The IRX1 gene is a member of the iroquois homeobox gene family and plays a role during pattern formation of vertebrate embryos. In published literature, there are only four studies investigating the effects of epigenetically silencing IRX1 gene, in the patients with gastric cancer (33) and the studies of Bennett et al with HNSCC (7,8,34). Bennett et al (7) reported an association between the HNSCC and hypermethylation of IRX1 in accordance with our results, and four other genes (FUSSEL18, EBF3, SLC5A8 and SEPT9), but not SLC5A8; we did not find any methylated CpG on the promoter regions of the last five genes. In two other studies, Bennett et al also showed the association between IRX1 methylation and recurrence and between four other genes and clinicopathological parameters such as HPV status, alcohol and tobacco usage (8) and they investigated the interactions with some molecules of the TGF-β pathway, and their transcriptional inactivation by methylation in HNSCC caused the decrease in apoptosis and differentiation, and increased proliferation (34). In the present study, we only observed the methylation of the IRX1 gene which was the one of the five frequently methylated genes identified by Restriction Landmark Genomic Scanning (RLGS) method in metastatic HNSCC samples compared to primary tumors. Mass array methylation (7) and COBRA (8) analysis were used in fresh/paraffin-embeded HNSCC tumors and matched normal mucosa samples in the previous studies of Bennett et al (7,8). We screened the methylated CpG islands on the promoter regions of these genes by bisulfite sequencing using frozen HNSCC tumors, HNSCC cell lines and normal salivary rinses in order to select the genes that have the methylated CpG islands on their promoter regions in tumor samples but no methylation in normal mucosal and normal salivary rinse samples. Then, the methylation levels were measured by QMSP technique in 33 HNSCC tumors, 16 salivary rinses and 14 normal mucosa as reported in our previous study (35). Our aim was to find a cancer-specific biomarker and to detect the tumor tissues in salivary rinses of the patients earlier and to screen the normal population. In addition, in a previous study (34), tonsillar carcinoma samples were used, whereas we quantified the methylation in the primary tumors of the oral cavity, oropharynx, hypopharynx, larynx, maxillary sinus, nasal floor and unknown primary/neck. Therefore, discrepant observations between these studies by Bennett et al (7,8,34) and our data may be due to confounding factors including anatomic site, methodology and sampling method. Therefore, the four other genes described by Bennett et al (7,8), may not be good biomarker candidates in salivary rinse samples despite being differentially methylated in primary tumor samples. The present study is the first to investigate the methylation levels of NOL4 gene and to show high methylation in the patients with HNSCC by QMSP technique.
In addition, it would be helpful to increase sample size to facilitate a more precise determination of accuracy of these biomarkers in detection of the disease and to evaluate the association of the results with the clinical parameters of the disease. These newly identified silenced genes here remain to be tested in saliva, serum or plasma samples from HNSCC patients in a larger sample size.