Expression profiles of selected genes in tumors and matched surgical margins in oral cavity cancer: Do we have to pay attention to the molecular analysis of the surgical margins?

Results. SFRP1 gene expression was statistically significantly lower in the tumor samples than in the surgical margins (0.30 ±0.36 vs 0.62 ±0.36; p < 0.01). No correlation was found between gene expression and clinical parameters, except DAPK1, where low expression correlated with alcohol abuse (0.85 ±1.19 vs 1.97 ±3.22; p = 0.074). Moreover, patients with G3 grade tumors, i.e., poorly differentiated tumors, had significantly higher values of DAPK1 gene expression than the G1 (well-differentiated tumors) and G2 (moderately differentiated) groups.


Introduction
Head and neck squamous cell carcinomas (HNSCCs) account for 3% of all diagnosed malignant tumors in men and 2% of those diagnosed in women. 1 Head and neck cancer is associated with an interplay between genetics and environment.It is associated with abnormalities in proliferation, apoptosis, differentiation, cell cycle regulation, DNA repair, signal transduction, and angiogenesis. 2The instability of the genome, chromosomal rearrangement and loss of DNA fragments are associated with carcinogenesis of HNSCC. 3 Califano et al. proposed a model for the dynamics of chromosomal damage in the course of cancers of the head and neck. 4Through progression, an increased amount of chromosomal loss takes places.The conversion from normal mucosa to invasive cancer is linked with an accumulation of genetic changes in tumor suppressor genes and proto-oncogenes.Furthermore, adjacent areas share similar genetic alterations.For HNSCC, the list of environmental risk factors includes alcohol consumption, tobacco use, poor oral hygiene, chronic exposure to certain industrial agents, and infection with specific subtypes of human papillomavirus (HPV). 1 Slaughter et al. first proposed the idea of "field cancerization" as the "preconditioning of an area of the epithelium to cancer growth by a carcinogenic agent", which means that an area with genetically altered cells could play a crucial role in the carcinogenic process based on many molecular findings. 5,6The initiation of field cancerization is associated with various types of molecular damage, e.g., altered gene expression.The residual field in the region adjacent to the tumor can cause local recurrences and 2 nd primary tumors after surgical intervention for the primary carcinoma.Actually, there are many theories interpreting oral field cancerization. 7The "cancer field effect" has been explained by the presence of cells with genetic alterations; however, involvement of epigenetic alterations in field cancerization has also been shown.Epigenetic changes are defined as alterations in gene expression that are not coded in the DNA sequence.Among epigenetic mechanisms, such as DNA methylation at CpG sites or histone modifications, aberrant DNA methylation has been frequently analyzed in various types of cancer.Hypomethylation leads to the activation of cancer-associated genes, whereas hypermethylation of promoter CpG islands is associated with the silencing of various tumor-suppressor genes.Several environmental factors could induce epigenetic modifications. 8In this study, we have analyzed the gene expression of TIMP3, SFRP1, SFRP2, CDH1, RASSF1, RORA, and DAPK1 in primary HNSCC tumors and matching surgical margin samples.We selected genes involved in degradation of the extracellular matrix, cellular proliferation, migration, and apoptosis.Disruption of these processes can lead to carcinogenesis.][11][12][13][14][15] The main aim of the study was to provide more information concerning the molecular mechanism of oral malignancy based on gene expression, which could provide valuable information for a better understanding of the oral carcinogenesis process.

Patients and tissue samples
We collected 56 primary HNSCC tumors and 56 matching surgical margin samples following surgical resection from patients at the Clinic of Oncological and Reconstructive Surgery of Maria Skłodowska-Curie Memorial Cancer Center and the Institute ).An informed consent form was obtained from all patients.None of the patients included in this study had preoperative radio-or chemotherapy.The average age of the patients was 56.05 years (range: 29-77 years, median: 58.5 years).There were 37 men (66%) and 19 women (34%); 80% of the patients (45/56) were smokers; 73% of them (41/56) reported alcohol consumption; 64% (36/56) were both smokers and alcohol consumers.Tumor staging was based on the pathology findings and then revised according to the 2007 version of the American Joint Committee on Cancer (AJCC) staging system for analysis. 16,17The clinical parameters of the patients are shown in Table 2.

Homogenization and ribonucleic acid extraction
The tissue samples were fixed in RNAlater ® Stabilisator Reagent (Sigma-Aldrich, Saint Louis, USA) and frozen to -80°C, then thawed slowly at room temperature and homogenized with FastPrep ® -24 homogenizer (MP Biomedicals, Santa Ana, USA) using tubes with ceramic beads for tissue homogenization, Lysing Matrix D (MP Biomedicals, Santa Ana, USA).The RNA was isolated using an RNeasy ® Mini Kit (Qiagen, Hilden, Germany).In addition to the standard procedure, DNase I digestion was performed on the extracted RNA (RNase Free DNase Set, Qiagen, Hilden, Germany) to remove the residual genomic DNA.
The RNA was quantified by measuring the UV absorbance at 260/280 nm (NanoDrop ™ ND, 1000 Spectrophotometer, Thermo Fisher Scientific, Waltham, USA) and the integrity was assessed by electrophoresis on 1.2% agarose gel stained with ethidium bromide (Serva, Heidelberg, Germany).Additionally, the RNA integrity number (RIN) was derived with an Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, USA) using an Agilent RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, USA); this helped to ensure RNA quality.

Complementary DNA (cDNA) synthesis
The total RNA from each tumor and surgical margin sample was reverse-transcribed into cDNA using a High Capacity cDNA Reverse Transcription Kit with RNase Inhibitor (Applied Biosystems, Foster City, USA).The Total RNA (30 ng) was reverse-transcribed into cDNA.The reaction was performed with a volume of 20 µL according to the manufacturer's instructions and using Mastercycler ® Personal (Eppendorf, Hamburg, Germany).To avoid multiple thawing, cDNA samples were divided into a number of portions, which were sufficient for all subsequent quantitative polymerase chain reactions (Q-PCR); these portions were frozen at −80°C.
The Q-PCR for all genes was performed in a volume of 20 µL using 1 µL of cDNA, 10 µL of TaqMan ® Gene Expression Master Mix (Applied Biosystems, Foster City, USA), 1 µL of primer and probe mix (TaqMan ® Gene Expression Assays), and 8 µL of H 2 O (Qiagen, Hilden, Germany).All assays were carried out in 96-well plates (Applied Biosystems, Foster City, USA) on a 7300 Real-Time PCR System and were analyzed by SDS v. 1.4 software (Applied Biosystems, Foster City, USA).The glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH, Hs99999905_m1) was used as an internal control.The expression levels of all these genes were normalized to those of GAPDH.In all amplification reactions, a negative control was also included: water free of DNase, RNase, and protease (5Prime, Hilden, Germany) was used in place of cDNA.Thermal cycling for all analyzed genes and GAPDH was 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and at 60°C for 1 min.The comparative threshold cycle (Ct) method 2 -∆∆Ct was used to determine the relative gene expression levels (relative quantification -RQ) for each of the 7 target genes.Five surgical margin samples were used as a calibrator.Relative mRNA expression was determined from the tumor and surgical margin samples using mRNA expression from the calibrator.

Statistical analysis
Statistical analysis was performed using STATISTICA v. 10.0 PL (QUEST, Tulsa, USA).Statistical significance was set at a p-value of less than 0.05.All tests were twotailed.Imputations were not performed for missing data.Nominal and ordinal data were expressed as percentages, while interval data were expressed as means ±standard deviation.The distribution of variables was evaluated by the Shapiro-Wilk test and the homogeneity of variances was assessed by the Levene test.For comparison of data between 2 groups, the Student's t-test was used for independent data.For comparison between different histological grades, one-way analysis of variance (ANOVA) was used with an LSD post-hoc test.

Results
When gene expression levels were compared between the tumor samples and the margin samples, a statistically significantly lower gene expression of SFRP1 was found in the tumor samples.The RQ values of the selected genes are reported in Table 3.
We also analyzed the correlation of the clinical parameters with the expression levels of selected genes.No association was found between the gene expression levels and clinical parameters, except DAPK1, in which a low gene expression level statistically correlated with alcohol abuse (Table 4).Moreover, the one-way ANOVA showed a significant influence of the histological stage.The LSD post-hoc test showed that the patients with high-grade G3 tumors -i.e., poorly differentiated -had a significantly higher gene expression level of DAPK1 than the patients with low-grade G1 (well-differentiated) tumors (p < 0.05) or G2 (moderately differentiated) tumors (p < 0.01).These results are shown in Fig. 1.

Discussion
28,29 Simultaneous analysis of gene expression levels in the tumor and surgical margins has rarely been performed. 23,26,30Cancer development is the result of the accumulation of genetic and epigenetic alterations.Because changes in certain genes occur in the very early stages of tumorigenesis, the detection of preneoplastic alterations in the surgical margin could facilitate the detection of cancer.A molecular approach to the matching margin can contribute to cancer prevention and control; therefore, we studied the tumors and matched surgical margins from the patients.Moreover, in our study, the expression levels of the TIMP3, SFRP1, SFRP2, CDH1, RASSF1, and RORA genes did not

G3
correlate with any of the clinical-pathological parameters, whereas DAPK1 correlated with both the histological grade of the tumor and alcohol consumption.8][31][32][33]

TIMP3
The expression of TIMPs in oral squamous cell carcinoma (OSCC) shows a trend that is higher in tumors than in normal tissue, and which correlates with an increased risk of metastasis and regional lymph node involvement, although some deviations from this have been noted. 34here are suggestions that TIMP3 protects against tumor development by suppressing angiogenesis, tumor growth and metastasis through inferred apoptosis. 9oskunpinar et al. analyzed the expression levels of 88 genes in laryngeal carcinoma using a PCR array and showed an altered expression of 16 genes when compared to paired cancer-free tissue. 30One of the altered genes in which expression was significantly higher in the tumor tissue was the TIMP1 gene.TIMP1 protein expression was significantly higher in laryngeal squamous cell carcinoma than in adjacent non-cancerous tissues. 19Similarly, an increased expression of TIMP2 has been observed in tumor tissues compared with normal tissues. 26Burduk et al. described oropharyngeal cancer without lymph node metastasis showing lower TIMP1 and TIMP2 protein expression in cancer cells and stroma compared to patients with lymph node metastasis. 35On the other hand, it has been reported that a downregulated expression of the TIMP3 protein in esophageal adenocarcinoma is associated with enhanced tumor invasiveness and reduced patient survival 9 and that the downregulation of TIMP3 stimulates growth and invasion in endometrial cancer cell lines. 36The present study of TIMP3 gene expression in tumors and surgical margin tissues and the correlation between gene expression and clinicalpathological parameters in oral cavity cancers indicated no significant differences (Table 3).

SFRP1 and SFRP2
8][39] In our study, we examined SFRP1 and SFRP2 gene expression levels and found a statistically significantly lower expression of SFRP1 in tumor samples compared to margin samples, although the difference was not significant for the SFRP2 gene (Table 3).Sogabe et al. proved that in OSCC cell lines, the silencing of SFRP genes and their loss of function lead to cell proliferation during oral carcinogenesis. 38imilarly, Xiao et al. showed that SFRP2 mRNA expression was downregulated in tumor samples of OSCC compared to tumor-adjacent normal tissue, and that the loss of expression was connected with hypermethylation of the gene promoter. 39Reduced SFRP1 expression was also detected by immunohistochemical staining in a group of patients diagnosed with mucoepidermoid carcinoma of the salivary glands. 40The SFRP1 protein can act as a Wnt signaling inhibitor by attaching extracellular Wnt ligands or directly binding to the transmembranous receptor FZD.Its role as a potential progression marker was clarified in a study by Rogler et al. which used cell cultures and tumor samples of bladder cancer. 37They demonstrated that the function of SFRP1 in the process of oncogenesis is more complicated, considering the non-canonical Wnt-and mitogenactivated protein kinase (MAPK) signaling pathways.

CDH1
A dysfunction of cadherin 1 is involved in carcinogenesis, and a loss of function has been demonstrated to promote tumor invasion and metastasis in different cancer models. 41t has been reported that the loss of protein expression of CDH1 is also associated with an increased invasive potential in head and neck cancer. 42In a study conducted on clinical samples collected from patients with tongue squamous cell carcinoma, a significantly lower CDH1 mRNA expression level than in the corresponding noncancerous tissues was shown. 31In our study, no significant differences were found between the CDH1 gene expression levels of the tumors and the surgical margin samples.Similar results were obtained by Kordi-Tamandani et al., who found no significant differences between the mRNA expression of the CDH1 gene of patients with oral cavity cancer and that of a healthy control group. 12

RASSF1
Another gene that plays an important role in human cancer cell growth and progression is RASSF1, and an abnormal expression of RASSF1 could be an important step in oncogenesis. 21In the case of the RASSF1 gene, we found no significant differences in our study.A downregulated expression of RASSF1A transcripts and protein in tumor tissues in esophageal and nasopharyngeal carcinomas were observed by Lo et al.; moreover, a reduced expression correlated with the histological grade of the tumor. 33The mRNA expression of RASSF1A was also downregulated in primary tumors in a group of patients with cutaneous melanoma and lacrimal gland carcinoma compared to healthy groups, and was also downregulated in lung and breast cancer cell lines. 18,21,43Furthermore, aberrant methylation of RASSF1A has been observed in several cancer types 43 and a few reports have been focused on the methylation of this gene in HNSCC. 44,45It is well known that hypermethylation is one of the important epigenetic mechanisms responsible for inactivation of the gene, in addition to genetic alterations of gains and losses. 12,43

RORA
RORA is a steroid hormone receptor involved in cellular processes, including metabolism, angiogenesis, inflammation, and the circadian rhythm.There are several diseases where RORs are integral to the onset and progression, such as autoimmune diseases, inflammation, osteoporosis, and cancer, and it has been proven that in tumor cell lines and cancerous tissues, ROR mRNA levels are often downregulated. 46RORA expression was analyzed in colorectal adenocarcinomas, where RORA mRNA expression was downregulated in comparison to normal colonic tissue. 47n this study, we found no significant differences in RORA gene expression, and to our knowledge there have been insufficient studies on the expression of this gene in patients with head and neck cancer.Sørensen et al. analyzed gene expression in human squamous cell carcinoma cell lines and observed a higher level of RORA gene expression in the hypoxia group compared to the control group; this was one of the genes induced at low oxygen independent of pH. 29Genes upregulated by low oxygen have been considered endogenous markers for tumor hypoxia.

DAPK1
DAPK1 is a regulator of apoptosis; suppression of this gene is thought to be critical in the development of tumors.Lower mRNA and protein expression of DAPK1 was observed in a group of patients with primary gastric cancer compared with adjacent non-tumor tissues. 48Mariano et al. detected that even with losses of copy numbers for the DAPK1 gene, the immunohistochemical reaction showed protein expression of this gene in a group of patients with salivary gland neoplasms. 49Our study showed no significant differences in DAPK1 gene expression between the tumors and the surgical margin tissues, but there was a significant association between DAPK1 gene expression and tumor grade.Patients with G3 tumors had significantly higher RQ values of DAPK1 than patients with G1 and G2 tumors.Figure 1 shows the histological findings of the tumors as related to the gene expression of DAPK1.It is often difficult to compare findings between studies because of the different populations and methods used, but our results regarding DAPK1 are not compatible with other studies.Another study revealed that the methylation of the DAPK1 gene was associated with the progression of HNSCC. 44,45Aberrant promoter DNA methylation of this gene has been examined in other types of cancer, including breast cancer; furthermore, tumors with an advanced T-category revealed a higher frequency of DAPK1 methylation. 50Surprisingly, the data of Brait et al. on DAPK1 promoter methylation showed the frequencies of methylation for DAPK1 in normal thyroid samples to be higher than the frequencies in cancer samples. 51According to the research, these results limit the usefulness of this gene as a diagnostic marker; additionally, the hypermethylation in the tumors in neoplastic relevance is questionable.Our understanding of the physiological role of the gene is still at an early stage.Furthermore, our studies of the selected tumor histological grade were done on a small population; we must carry out further studies on a larger population in order to verify this result.
Interestingly, we also found a lower DAPK1 gene expression in the group of patients with alcohol abuse (0.85 vs 1.97; Table 4).This finding could be explained by an epigenetic mechanism after exposure to a risk factor like alcohol consumption, as different lifestyle factors induce expression changes in different genes.Also, low transcription could be associated with methylation induction, as genes specific to a factor could be methylated. 8n the development of upper aerodigestive tract cancer, alcohol and tobacco are 2 well-established associated risk factors. 52here are many different reasons and hypotheses for altered gene expression in tumors and surgical margin tissue.A tumor's heterogeneity and microenvironment are undoubtedly linked to the biology of HNSCC.The clinical application of genetic alterations and their role in HNSCC progression are still being discussed. 41Different mutations of the TP53 gene are most prevalent among the numerous observed mutations in HNSCC.An association between the specific mutations of this gene and the biological and clinical course of cancer has been found. 53Moreover, epigenetic events, such as aberrant promoter gene hypermethylation, are often observed not only in tumor tissue, but also in surgical margins. 39,54,55Abnormal DNA methylation patterns in promoter regions can inactivate genes and facilitate tumor formation and progression. 56nvironmental factors like exposure to alcohol and cigarettes can influence aberrant methylation patterns, too. 57part from tobacco and alcohol consumption, HPV has been named as a causative agent in a subset of this cancer. 58HPV infection leads to deregulation of the cell cycle and it is well-known that additional genetic changes are needed for HPV-mediated oncogenesis. 59It was found that an overexpression of p16 was connected with molecular events occurring after HPV infection and p16 has been used as a surrogate marker for evaluation of HPV status. 60,61Interestingly, HPV was also found to deregulate the methylation levels in individuals with HPV infection. 62ecently, Helicobacter pylori was detected in samples collected from malignancies in the oropharyngeal area and its influence on carcinogenesis was also suggested. 63t is essential to further study the methylation status in this group of patients, and papers on this subject are currently in progress.One limitation of this study was the fact that we did not investigate the patients' HPV status; we are also considering further investigation of this issue.
In our study population, 80% of the patients were current smokers (most of them smoked more than 1 pack per day); 73% of them reported alcohol consumption, and 64% both smoked and consumed alcohol.Given the important role that environmental factors such as alcohol abuse and tobacco exposure play in the onset of cancer, it is clear that some genetic or epigenetic alteration occurs in noncancerous tissues adjacent to the tumor tissue. 1 As many of the patients studied were exposed to these factors, they showed altered expression levels, and the lack of statistically significant differences between the tumor and the margin can also be interpreted in terms of "field cancerization."It is known that patients with head and neck cancer subsequently have increased morbidity and mortality and that subsequent primary tumors are the main reason for mortality in this group of patients. 64It is wellknown that the carcinogenic process involves a progressive accumulation of genetic abnormalities and that HNSCC is a diverse disease with complex molecular abnormalities.A special model of genetic alterations and the progression of squamous dysplasia to invasive carcinoma has been described, including the clonal growth of transformed cells with the formation of an abnormal genetic field. 65Moreover, oral field cancerization suggests that oral cancer does not arise as an isolated cellular phenomenon but rather as an anaplastic tendency involving many cells at once and it results in the multifocal development of cancer at varied rates within the total field as a reaction to a carcinogen, particularly tobacco.This concept could explain the appearance of multiple primary cancers and recurrences despite the total excision of oral cancer. 7Tabor et al. found genetically altered fields in the non-neoplastic mucosa surrounding the tumors of 10 out of 28 patients (36%) with HNSCC -the size of the fields was variable.Moreover, in 7 out of 10 patients the field extended beyond the surgical margin. 66n order to understand specific tumor behavior and the microenvironment, further studies are needed.As a lot of clinical procedures are limited predictors of tumor progression, many authors suggest that the detection of abnormalities in the field defect can be a useful diagnostic marker to help in assessing the risk of cancer. 67,68It would be favorable to change the assessment of a safe margin extension or to expand the irradiation field. 69The discovery of a marker of cancerization is important, but the oncogenesis process is very complicated and molecular techniques have limitations as well.Finding a marker for the development of cancer which provides insight into its earliest stages requires attention to also be focused on a molecular analysis of the surgical margin.

Conclusions
To find markers connected with cancer development and to provide insight into the earliest stages of cancer development, attention should be focused on a molecular analysis of the surgical margins.More investigation is required to completely understand all of the components.

Fig. 1 .
Fig. 1.Relative quantification (RQ) values of DAPK1 in the group of patients with HNSCC according to G1, G2 and G3 grading stage

Table 1 .
The characteristics of selected genes This gene encodes RORA.The protein encoded by this gene is a member of the NR1 subfamily of nuclear hormone receptors.These receptors are critical in the regulation of a number of physiological processes.

Table 2 .
Clinical features of patients

Table 3 .
Relative quantification (RQ) values in tumor vs surgical margin in patients with HNSCC

Table 4 .
Relative quantification (RQ) values of DAPK1 in the group of patients with HNSCC with/without alcohol abuse