Introduction

Oral squamous cell carcinoma (OSCC), the incidence of which is increasing, is the most common type of malignant tumor of the oral cavity. Late-stage presentation of OSCC is common [1]. Recently, changes in the abundance of oral microbiota have been associated with human oral cancers [2]. Although the etiopathogenesis of OSCC appears to be complex, with interactions between genetic, environmental, lifestyle, and other factors, it is generally considered to be an immunologically mediated process characterized by the appearance of an intense band-like leukocyte infiltrate at the epithelium-connective tissue interface [3].

Immunological imbalances created by infiltrating inflammatory cells may contribute to the growth [4] and spread of oral cancer. The cytokine networks associated with several common tumors are rich in inflammatory cytokines, growth factors, and chemokines [5]. In particular, inflammatory cytokines can be produced directly by tumor cells and/or tumor-associated leucocytes and platelets that may contribute directly to malignant progression [6]. Local cytokine production within the tumor microenvironment can prevent the effector cell response [7], and cytokines can also mediate the activities of immune cells against malignant cells [8].

In humans, the cytokine interferon-gamma (IFN-γ; also known as type II interferon) is critical to both the innate and adaptive immunity [9]. IFN-γ, the production of which is related to the induction of T lymphocyte and macrophage reactivity, maintains the major histocompatibility class II molecule expression, thus participating in keratinocyte apoptosis and oral lichen planus (OLP) disease chronicity [10] and further enhancing immune responses against malignant cells. Given the importance of IFN-γ in human immune responses, it is unsurprising that genetic and epigenetic variations within IFN-γ are associated with a range of diseases, including cancer [11].

Changes in DNA methylation patterns, particularly in the promoter regions of genes, can have profound effects on gene expression and have accordingly been implicated in the development of periodontal disease [12]. Methylation profiling of cancer cells and studies of individual genes have disclosed the frequent occurrence of gene-specific hypermethylation in diverse cancers (e.g., breast cancer [13] and cervical cancer [14]). Although CpG methylation of the IFN-γ promoter is considered a negative transcriptional regulatory event that affects IFN-γ production [15], such events have not been investigated in oral cancer tissues. Therefore, the present study aimed to evaluate the methylation status of the IFN-γ promoter region in oral cancer tissues compared with that in normal and oral epithelial dysplasia tissues.

Materials and methods

Clinical samples and DNA extraction

The archived tissue samples from 85 cases of OSCC, 47 cases of dysplastic lesions of the oral cavity, and 53 normal biopsies were utilized in this study. Clinical characterization of the OSCC patients is summarized in Table 1. Oral tissues were collected via oral mucosa biopsy, flash-frozen in liquid nitrogen, and stored at −150 °C. All patients provided informed consent prior to sample collection according to institutional guidelines. This protocol was approved by the ethics committee of the Second Hospital of Hebei Medical University, Shijiazhuang, Hebei Province, China.

Table 1 The methylation of the IFN-γ gene promoter in normal oral tissue, dysplasia, and OSCC
Full size table

Following the detection of malignancy, patients underwent surgical resection of primary oral cancers at the Department of Oral Surgery, Second Hospital of Hebei Medical University. The histological type and grade of each tumor was classified according to the TNM criteria set forth by the Union for International Cancer Control/American Joint Committee on Cancer, 7th edition. Tissues from samples diagnosed as dysplasia and tumorigenesis were obtained via microexcision. All primary tumor tissues and control samples were diagnosed following hematoxylin and eosin (HE) staining. Frozen tissue samples were subjected to genomic DNA extraction using standard proteinase K treatment followed by phenol/chloroform extraction. DNA concentrations were determined with a spectrophotometer.

Bisulfite modification

Tissue methylation patterns were assessed using bisulfate-induced DNA modification in a manner similar to that reported by Goldenberg et al. [16]. In brief, bisulfite treatment converts unmethylated cytosines in DNA to uracil, whereas methylated cytosines remain unmodified. Extracted genomic DNA was modified using a bisulfite conversion kit (EZ DNA Methylation-Gold™ Kit; Zymo Research Corp.). Modified DNA was ready for immediate analysis or was stored at or below −20 °C for later use or at or below −80 °C for long-term storage. Each polymerase chain reaction (PCR) included 1 μl of eluted DNA.

Methylation-specific PCR

Methylation-specific PCR (MSP) uses specific primers for methylated or unmethylated DNA to distinguish the presence of methylation based on alterations produced after bisulfite treatment. All samples in this study were analyzed by MSP. The primer sequences for methylated and unmethylated IFN-γ were described previously [14]. PCR was carried out in a total volume of 20 μl which contained 2 μl of bisulfite-treated genomic DNA, 1× PCR buffer, 0.25 μM of each primer, 250 μM (each) dNTP mix, and 0.75 unit of FastStart Taq DNA polymerase (Roche Applied Science). Thermal cycling was initiated at 95 °C for 5 min, followed by 40 cycles of 95 °C for 30 s; specific annealing temperature for 30 s; extension temperature at 72 °C for 30 s; and a final extension at 72 °C for 2 min. An untreated blood DNA from a normal individual was used as a negative control. A methylation-positive DNA control was made in vitro using SssI methylase (New England Biolabs, Beverly, MA, USA) which methylated every cytosine of CpG dinucleotide in the DNA. Water blanks were included with each assay. The same PCR conditions were used for tumor, dysplasia, and normal tissue DNA. PCR products were visualized on 2% agarose gels stained with ethidium bromide. Positive amplification only of unmethylated primers was interpreted as unmethylation. Positive amplification only of methylated primers or both methylated and unmethylated primers was defined as methylation.

Direct sequencing

Methylated and unmethylated PCR products were confirmed by direct sequencing. PCR products were gel purified and analyzed on an automated DNA sequence analyzer (ABI 3730xl, Life Technologies, Carlsbad, CA, USA) with a BigDye® Terminator kit (Life Technologies).

IFN-γ messenger RNA determination

Total RNA was extracted from the OSCC, dysplastic lesions, and normal tissues, which were homogenized with gentle MACS™ Dissociator (Miltenyi Biotec) using the TRIzol (Invitrogen) according to the manufacturer’s instructions. The quality of the RNA was determined using a BioSpectrometer (Eppendorf). One microgram of RNA was subjected to reverse transcription using a first-strand cDNA synthesis kit (Invitrogen) according to the manufacturer’s instructions. RT-qPCR of messenger RNAs (mRNAs) was performed using a Platinum SYBR Green qPCR Super Mix UDG Kit (Invitrogen), and real-time PCR experiments were carried on an ABI 7500 FAST system (Life Technologies). Primer sequences for IFN-γ were described previously [14]. A relative amount of transcripts was normalized with β-actin and calculated using the 2−ΔΔCt formula. All real-time RT-PCRs were performed in triplicate.

Immunohistochemistry

Tissue samples were dehydrated in an ethanol series, cleared in xylene, and embedded in paraffin. Four-micrometer r sections were prepared and mounted on poly-l-lysine-coated slides. Immunohistochemical analysis was done using a commercially available kit (Invitrogen, Carlsbad, CA, USA). Sections were incubated at 60 °C for 30 min and deparaffinized in xylene. Endogenous peroxidase activity was quenched by incubation in a 9:1 methanol/30% hydrogen peroxide solution for 10 min at room temperature. Sections were rehydrated in PBS (pH 7.4) for 10 min at room temperature. Sections were then blocked with 10% normal serum for 10 min at room temperature followed by incubation with anti-IFN-γ antibodies (Proteintech) at a dilution of 1:100 for 16 h at room temperature. After washing thrice in PBS, the sections were incubated with secondary antibody conjugated to biotin for 10 min at room temperature. After additional washing in PBS, the sections were incubated with streptavidin-conjugated horseradish peroxidase enzyme for 10 min at room temperature. Following final washes in PBS, antigen-antibody complexes were detected by incubation with a hydrogen peroxide substrate solution containing a aminoethylcarbazole chromogen reagent. Slides were rinsed in distilled water, coverslipped using aqueous mounting medium, and allowed to dry at room temperature. The relative intensities of the completed immunohistochemical reactions were evaluated using light microscopy by three independent, trained observers who were unaware of the clinical data. All areas of tumor cells within each section were analyzed. All tumor cells in ten random high-power fields were counted. A scale of 0 to 3 was used to score relative intensity, with 0 corresponding to no detectable immunoreactivity and 1, 2, and 3 equivalent to low, moderate, and high staining, respectively.

Immunofluorescence

Immunofluorescence staining was performed with 4-μm paraffin cross sections from the OSCC tissues (n = 85), dysplastic lesions of the oral tongue tissues (n = 47), and normal tongue tissues (n = 53). After deparaffinized with xylene and rehydrated, the slides were pre-incubated with 10% normal goat serum (710027, KPL) and then incubated with primary mouse monoantibody anti-MAC2 (60207-1, Proteintech) and rabbit polyantibody anti-IFN-γ (18013–1, Proteintech). Secondary antibodies were a fluorescein-labeled antibody to mouse IgG (021815, KPL) and a rhodamine-labeled antibody to rabbit IgG (031506, KPL). In each experiment, DAPI (157574, MB Biomedical) was used for nuclear counter staining. Images were captured by confocal microscopy (DM6000 CFS, Leica) and processed by LAS AF software.

Enzyme-linked immunosorbent assay

Enzyme-linked immunosorbent assay (ELISA) (Total Survivin TiterZyme® Enzyme Immunometric Assay; Assay Designs, Ann Arbor, MI, USA) was used to determine the IFN-γ levels in tissue protein extracts of samples from patients with OSCC (n = 85), dysplasia (n = 47), and control tissues (n = 53). The undiluted samples were loaded in duplicate onto a 96-microtiter plate coated with antihuman IFN-γ monoclonal antibody. To quantify the IFN-γ concentrations, a standard curve of increasing quantities of purified recombinant IFN-γ was created for each experiment. After incubation at room temperature on a plate shaker for 1 h at 500 rpm, excess sample or standard was washed out, and a rabbit polyclonal antibody specific for IFN-γ was added. After another 1-h incubation, the excess antibody was washed out and a HRP-conjugated goat anti-rabbit IgG antibody was added. The excess conjugate was washed out after a 30-min incubation, after which 3,3′,5,5′-tetramethylbenzidine substrate solution was added to induce the colorimetric reaction. After another 30-min incubation, the enzymatic reaction was stopped and the plate absorbance was immediately read at 450 nm. The measured optical densities were directly proportional to the concentrations of IFN-γ for both the standards and samples.

Statistical analysis

Data are presented as means ± standard errors of the mean. One-way ANOVA was used to assess the association of gene expression with histopathological parameters. Student’s t test was used to compare the parametric data between groups. An analysis of variance and Bonferroni’s post test were used to compare multiple groups. A P value <0.05 was considered statistically significant. All statistical analyses were performed with SPSS 13.0 (SPSS, Inc., Chicago, IL, USA).

Results

IFN-γ mRNA and protein levels correlate with oral tumorigenesis

Although RT-PCR detected IFN-γ mRNA expression in control, dysplasia, and OSCC tissues, the expression was obviously decreased in oral cancer tissues (Fig. 1a). IFN-γ staining was markedly decreased in OSCC tissues versus controls and dysplasia tissues via immunohistochemistry (Fig. 1b). ELISA was used to quantify IFN-γ protein expression in control, dysplasia, and OSCC tissues. Notably, the IFN-γ protein level was modestly increased in dysplasia tissues compared with control tissues, whereas a more robust reduction was observed in oral cancer tissues, consistent with the results of the PCR analysis (Fig. 1b, c).

Fig. 1
figure 1

Oral tumorigenic tissues express relatively low levels of IFN-γ mRNA and protein. a Real-time reverse transcription polymerase chain reaction demonstrated differences in IFN-γ mRNA expression levels between groups. *P < 0.05 and # P < 0.01 compared with dysplasia samples. b IFN-γ protein levels in control (n = 53), dysplasia (n = 47), and oral squamous cell carcinoma (OSCC) tissues (n = 85) were analyzed via enzyme-linked immunosorbent assay. c Paraffin-embedded tissue sections were prepared from matched pairs of human primary OSCC, dysplasia, and normal tissues and were subjected to immunohistochemistry with an IFN-γ-specific antibody. Representative images of IFN-γ staining in human normal tissue (a), dysplasia tissue (b), and primary OSCC (c) are shown. Scale bars = 50 μm

Full size image

The IFN-γ promoter region is frequently methylated in human oral carcinomas

Epigenetic changes in cytokine genes can affect the ability of a cell to express these genes [17]. In human tumors, aberrant promoter region methylation causes the loss of expression of many cytokine genes [18]. To determine whether the IFN-γ promoter is hypermethylated in human oral carcinomas under physiological conditions, we isolated genomic DNA from 85 primary tumor specimens derived from human patients with oral cancer, 47 dysplasia specimens derived from human patients with oral dysplasia, and 53 normal specimens. MSP analysis revealed methylation of the IFN-γ promoter in 55.3% (47/85) primary oral carcinoma specimens, whereas only 38.3% (18/47) of dysplasia specimens and 22.6% (12/53) of normal specimens exhibited methylation (shown in Table 2). Aberrant IFN-γ promoter methylation and representative bands from the MSP analysis of IFN-γ are presented in Fig. 2a, c.

Table 2 Relationship between methylation status of the IFN-γ gene and clinicopathological characteristics in OSCC tissues
Full size table
Fig. 2
figure 2

Methylation status polymerase chain reaction analysis of IFN-γ in human oral tissues. a IFN-γ methylation statuses of human primary oral squamous cell carcinoma (OSCC), dysplasia, and normal specimens. Representative bands are shown. P1–P3 were derived from 85 patients with oral cancer, P4–P7 from 47 patients with dysplasia, and P8–P10 from 53 normal controls. Genomic DNA was modified with sodium bisulfite as described in the “Materials and methods” section and analyzed via methylation status polymerase chain reaction with unmethylation (U) and methylation (M) primers. b Sequencing of IFN-γ with representative data indicating partial and total CpG methylation. c IFN-γ methylation rates in normal (n = 53), dysplasia (n = 47), and OSCC (n = 85) tissues

Full size image

DNA sequencing of MSP products

MSP products from the methylated DNA of control, dysplasia, and oral cancer samples were subjected to DNA sequencing, which yielded the expected nucleotide changes in control samples. The representative results of a bisulfite sequence analysis of the IFN-γ promoter are shown in Fig. 2b. One CpG site was observed in the amplicon after excluding the primer sites.

Transcriptional activation in response to IFN-γ methylation in oral tissues

Real-time RT-PCR was used to detect mRNA expressions and, thus, the effect of IFN-γ promoter methylation on transcriptional silencing in different groups. Oral tumors with methylation exhibited significantly downregulated (4.76-fold) IFN-γ mRNA expression compared with those without methylation (P < 0.01). Likewise, oral epithelial dysplasia tissues with methylation exhibited considerably downregulated mRNA expression compared with those without methylation (6.79-fold difference; P < 0.01; Fig. 3a, b).

Fig. 3
figure 3

IFN-γ CpG methylation and mRNA expression in oral tissues. IFN-γ mRNA expression was detected using quantitative real-time reverse transcription polymerase chain reaction (PCR) in oral dysplasia (a) and cancer tissues (b). M and UM indicate methylation and unmethylation, respectively. P < 0.01 versus UM

Full size image

IFN-γ expression in macrophages from oral tissues

Given that macrophages might comprise a main source of IFN-γ in tumors, we investigated the expression of IFN-γ in macrophages involved in oral cancers via dual immunofluorescence staining of human oral tissue sections to determine IFN-γ and Mac-2 expression. Our results showed a significant reduction in the mean percentage of IFN-γ + Mac-2-positive cells in oral cancer tissues compared with controls (1.3 ± 0.31 vs. 3.2 ± 0.11, P < 0.01). In contrast, we observed a significant increase in this cell population in oral epithelial dysplasia tissues (6.2 ± 1.1 vs. 3.2 ± 0.11 for controls, P < 0.05; Fig. 4a, b).

Fig. 4
figure 4

Coexpression of IFN-γ and Mac-2 in human oral cancer tissues. a Sections from human normal, dysplasia, and OSCC specimens were subjected to immunofluorescent staining. Red, green, and blue staining indicates IFN-γ, Mac-2, and DAPI (nuclei), respectively (scale bars = 50 μm). b Analysis of IFN-γ expression in Mac-2-positive cells from human oral cancer tissues as determined via immunofluorescence. Arrows indicate IFN-γ-positive macrophages. *P < 0.05 versus normal; ** P < 0.05 versus dysplasia

Full size image

Clinicopathological features and IFN-γ hypermethylation in oral cancer

The results of a methylation status-based multivariate analysis of the IFN-γ promoter regions in oral cancer tissues are shown in Table 2. The evaluated clinicopathological parameters included sex, age, clinical stage, histopathology grade, regional lymph node involvement, distant metastasis, primary tumor scale, muscle invasion, and depth of muscle invasion (tongue). IFN-γ methylation was found to associate significantly with the clinical stage (P = 0.021), histopathology grade (P = 0.014), and primary tumor (P = 0.024). However, IFN-γ methylation had no apparent associations with sex (P = 0.830), age (P = 0.827), regional lymph node involvement (P = 0.280), distant metastasis (P = 0.456), muscle invasion (P = 0.517), and depth of muscle invasion (P = 0.220).

Discussion

In tumor cells, gene expression is frequently regulated by epigenetic events [19]. In the present study, we evaluated the methylation patterns and expression of IFN-γ in histologically confirmed biopsies from patients with premalignant and malignant oral lesions. Our findings demonstrated hypermethylation of IFN-γ in most oral cancer tissues and, correspondingly, low levels of IFN-γ expression compared with control samples. In addition, we determined an association between oral carcinogenesis and DNA methylation in the IFN-γ promoter region.

To verify this hypothesized epigenetic regulation of IFN-γ production, gene transcription in each group was evaluated according to the MSP status. In the present study, we observed obvious decreases in IFN-γ mRNA and protein expression and a significant increase in IFN-γ methylation in oral cancer tissues. Notably, the methylated samples expressed lower levels of IFN-γ mRNA than the unmethylated samples in both the dysplasia and cancer groups. These data suggest that the decreased production of IFN-γ via the hypermethylation and silencing of its gene promoter promotes inflammation.

IFN-γ, which may control tumor cell immunogenicity via its selective production in the tumor microenvironment, has been shown to be a crucial component of the cancer immunoediting process [20]. Tumor-associated macrophages have been shown to exert both positive and negative effects on tumor growth in various types of cancer [21]. Therefore, delineation of the specific contributions of macrophages and related cytokines to oral tumorigenesis is a key to an understanding of oral cancer. Through our detection of IFN-γ expression in macrophages from human oral tissues, we observed a loss of IFN-γ expression in oral cancer tissues compared with control and dysplasia tissues, without a corresponding change in the number of macrophages. This suggests that future studies will be required to identify the biological effects of IFN-γ silencing by methylation and their relationship in macrophages to the malignant phenotype.

In this study, IFN-γ methylation did not associate with prognostic variables, such as sex, age at diagnosis, regional lymph node involvement, distant metastasis, muscle invasion (tongue), and depth of muscle invasion (tongue). However, IFN-γ methylation significantly correlated with the clinical stage (P = 0.021), histopathology grade (P = 0.014), and primary tumor (P = 0.024), indicating that IFN-γ exerts a significant influence on the growth and differentiation of both normal and malignant oral epithelia.

To summarize, hypermethylation-induced transcriptional silencing of IFN-γ in both premalignant and malignant oral lesions induces expression patterns distinct from those in the normal oral mucosa and suggests an important role for these changes in the progression from a premalignant state to malignancy. The detection and quantitation of promoter region methylation could facilitate clinical p T category of OSCC and contribute significantly to the screening, surveillance, and management of premalignant oral lesions and OSCC. Thus, aberrant IFN-γ promoter methylation could be related with tumorigenesis of oral cancer.