Introduction

Currently, oral cancer is the fifth most frequently occurring cancer worldwide and the fourth highest cause of male cancer mortality in Taiwan. More than 90 % of all oral malignancies are squamous cell carcinomas (SCCs) [1]. The high prevalence of oral cancer in Taiwan has been attributed to the high rate of people who habitually chew betel quid [2]. Approximately half of oral SCCs affect the tongue and floor of the mouth [3]. There are three main treatment procedures for oral cancer: surgery, chemotherapy, and radiation therapy. Despite advances in surgical techniques and adjuvant therapies, the 5-year overall survival rate for patients after a diagnosis of oral SCC remains poor [4]. The search for potential prognostic markers is still being studied.

The myeloid zinc finger 1 (MZF1) is a zinc finger transcription factor of the Kruppel family proteins originally cloned from the peripheral leukocytes of a patient with chronic myelogenous leukemia [5]. The gene encodes a 485-amino acid protein containing 13 C2H2 zinc fingers that are arranged in two distinct DNA-binding domains recognized as two independent DNA target sequences [6]. MZF1 plays a critical role in regulating the early stage of myeloid progenitor cell differentiation, including HL-60, KG1, HEL, and K562 human leukemia cells [5]. Transient or constitutive MZF1 expression inhibits hematopoietic development by downregulating both CD34 and c-myb promoter activity [7]. In addition, MZF1 transcriptionally regulates protein kinase C α expression in various human cancers cells, such as human hepatocellular carcinoma cells, breast cancer cells, and bladder transitional cell carcinoma cells [8–10]. Hsieh et al. have shown that treating the antisense oligodeoxynucleotides of MZF1 in human HCC cells inhibits cell migration, invasion, and tumor growth in nude mice [11]. In another study, overexpression of MZF1 induced metastasis in a solid tumor by increasing Axl promoter activity [12]. However, Tsai et al. indicated that MZF1 inhibits matrix metalloproteinase-2 transcription and reduces the invasiveness of human cervical cancer cells [13]. These observations suggest that MZF1 plays multiple roles in tumorigenesis, functioning as both a tumor promoter and tumor suppressor. However, the biological roles of MZF1 in OSCC remain poorly understood. In this study, we conducted an immunohistochemical analysis to investigate the relationships between the expression of MZF1 and clinicopathologic parameters in 274 patients with OSCC.

Materials and methods

Patients and tissue microarrays

We constructed formalin-fixed paraffin-embedded tissue microarrays composed of 274 OSCC tissue cores as previously described [14]. All patients with cancer underwent surgery from 2000 to 2006, which included wide excision of the primary tumor and neck dissection. Diagnosis of OSCC was based on histological examination of hematoxylin and eosin-stained tissue sections. Briefly, a slide with a representative tumor was selected and circled from each case, and the area corresponding to the selected area on the slide was also circled on the block. They were cored with a 1.5-mm-diameter needle and transferred to a recipient paraffin block. The recipient block was sectioned at 4 μm and transferred to silanized glass slides. Approval from the Institutional Review Board of Chung Shan Medical University Hospital was obtained prior to this study.

Immunohistochemical (IHC) analysis

Paraffin-embedded OSCC tissue sections (4 μm) of the paraffin slice on coated slides were washed with xylene to remove the paraffin and rehydrated through serial dilutions of alcohol followed by washings with a solution of phosphate-buffered saline (PBS, pH 7.2). After incubation with an anti-MZF1 (1:200 dilution; Santa Cruz Biotechnology, Santa Cruz, CA) antibody for 60 min at room temperature, slides were thoroughly washed three times with PBS. The conventional streptavidin peroxidase method (LSAB Kit K675; DAKO, Copenhagen, Denmark) was performed for signal development. The expression in the cytoplasm and nucleus of MZF1 was independently semiquantitatively assessed based on the intensity of staining by two pathologists who scored coded sections without knowledge of clinical and follow-up information. A final agreement was obtained for each score by using a multiheaded microscope (Olympus BX51 10 headed microscopes). The intensity of staining was respectively scored −, 1+, and 2+ for negative, weak, and strong staining.

Statistical analysis

Correlations of MZF1 with clinicopathologic parameters of OSCC were examined by Pearson’s χ 2 test. Cumulative survival was analyzed by the Kaplan-Meier method. Risk factors affecting survival were assessed by a Cox proportional hazards regression model using the SPSS statistical package (SPSS, Chicago, IL). A p value of <0.05 was considered statistically significant.

Results

Table 1 lists the clinicopathologic characteristics of patients with oral SCC. In this retrospective study, we enrolled 274 patients (259 men, 15 women) and analyzed their conditions. The patients were aged 31 to 90 years (mean age = 55.83 ± 11.14 years). The cancers were located at the following sites: buccal mucosa (n = 105), tongue (n = 91), gingiva (n = 35), palate (n = 16), floor of the mouth (n = 14), and others (n = 13). Among the patients, 18.2 % were at stage I (n = 50), 20.4 % were at stage II (n = 56), 12.4 % were at stage III (n = 34), and 48.9 % were at stage IV (n = 134). All patients were classified according to the seventh edition of the TNM staging system.

Table 1 Patient characteristics
Full size table

According to the level of nuclear MZF1 expression, we divided MZF1 immunohistological stains into two groups: positive (1+/2+) and negative (0) (Fig. 1). Nuclear MZF1 expression was present in 190/274 (69.3 %) cases. Our results revealed nonsignificant statistical differences between MZF1 nuclear expression and age, gender, cancer location, lymph node metastasis, distant metastasis, and grade. Patients with negative nuclear MZF1 expression were associated with more advanced clinical stages (p = 0.011) and a larger tumor size (p = 0.002) (Table 2). However, there were no significant statistical differences between MZF1 cytoplasmic expression and age, gender, cancer location, clinical stages, tumor size, lymph node metastasis, distant metastasis, and grade (Table 3).

Fig. 1
figure 1

Myeloid zinc finger 1 (MZF1) nuclear expression in oral cancer. Tissue microarrays of primary oral squamous cell carcinoma (OSCC) (274 cases) were immunohistochemically analyzed for MZF1. a, d No detectable MZF1 (0). b, e Weak expression levels (1+). c, f Strong expression levels (2+). g Negative control. h Positive control. a–c Low-power field (×100). d–h High-power field (×200)

Full size image
Table 2 Patient characteristics regarding nuclear MZF1 expression
Full size table
Table 3 Patient characteristics regarding cytoplasmic MZF1 expression
Full size table

Univariate analysis performed by Cox proportional hazards regression model identified advanced clinical stage (p = 0.013), large tumor size (p = 0.035), positive lymph node metastasis (p = 0.05), and negative nuclear MZF1 expression (p = 0.030) as correlating with poor survival for patients with oral tongue squamous cell carcinoma (Table 4).

Table 4 Univariate survival analyses of MZF1 and clinicopathological parameters among patients with oral cancer using the Cox proportional hazard regression model
Full size table

We performed the Kaplan-Meier analysis to evaluate the relationship between the expression of MZF1 and overall survival (Fig. 2). The results showed that oral tongue squamous cell carcinoma patients with negative nuclear MZF1 expression had a significantly lower survival rate (p = 0.028). The median survival in negative nuclear MZF1 expression was 28.8 months, whereas that in positive nuclear MZF1 expression was 78.6 months.

Fig. 2
figure 2

Kaplan-Meier survival curve showing the relation between nuclear MZF1 expression in primary tumors and survival in a 274 oral squamous cell carcinoma (OSCC) patients, b 105 buccal mucosa squamous cell carcinoma patients, and c 91 tongue squamous cell carcinoma patients. The overall survival of tongue squamous cell carcinoma patients with negative nuclear MZF1 staining was significantly lower than that of patients with positive nuclear MZF1 staining (p = 0.028, log-rank test)

Full size image

Discussion

Oral cancer has recently become a critical topic in Taiwan because the proportion of betel quid chewers is high [15]. Other risk factors for oral cancer include tobacco use, alcohol consumption, and human papilloma virus infection [16]. Oral cancer in Taiwan most frequently occurs on the tongue and buccal mucosa; cancer development at these sites is attributable to the habit of chewing betel quid [17, 18]. Tissue biopsies are performed to diagnose oral cancer, which is typically identified by advanced symptoms such as persistent bleeding. Despite advancements in therapy for patients with early-stage oral cancer, OSCC is still characterized as recurrent and involves a risk of tumor metastasis to cervical lymph nodes [19]. In this study, we evaluated MZF1 protein expression through immunohistochemistry, revealing that loss of nuclear MZF1 expression in patients with squamous cell carcinoma of the tongue is significantly associated with decreased overall survival rates.

MZF1 belongs to the Kruppel family of C2H2 zinc finger transcription factors, which are preferentially expressed in myeloid progenitor cells [5]. At least three isoforms of MZF1 proteins produced through alternative splicing have been reported [20, 21]. MZF1 has been shown to play a vital role in regulating gene transcription, repressing transcription in nonhematopoietic cells, and activating transcription in cells of hematopoietic origin [22, 23]. In addition, MZF1 and Sp1/Sp3 upregulate N-cadherin promoter activity, and N-cadherin controls the expression of phenotypic genes in osteoblasts [24]. Numerous reports indicated the relationship between MZF1 and tumorigenesis recently. Our results showed that negative nuclear expression of MZF1 was associated with advanced clinical stage and a larger tumor size in patients with OSCC (Table 2). Similar results were obtained when using Mzf1 −/− mice and determined that MZF1 negatively regulates the proliferative ability of hemopoietic cells [25]. Tsai et al. indicated that MZF1 binds to the promoter region of MMP2 and represses MMP2 transcription activity. The effect of MMP2 repression may be linked to inhibition of the migration potential of human cervical cancer cells [13]. In this study, we also found that loss of nuclear expression of MZF1 significantly related to poor overall survival in patients with oral tongue squamous cell carcinoma. These observations suggest that MZF1 functions as a negative regulator in tumorigenesis. However, on the contrary, some studies have demonstrated that MZF1 positively correlated with metastatic potential.

MZF1 was overexpressed in poorly differentiated human HCC cells and was essential for cell migration and invasion because it upregulated PKCα [8]. Mudduluru et al. revealed that MZF1 binds to the Axl promoter, transactivating promoter activity and promoting the metastatic potential of colorectal and cervical cancer cells [12]. In addition, MZF1 that binds to an ErB2-responsive enhancer element in the first intron of CTSB activates the signaling network of cysteine-cathepsin-mediated invasiveness [26]. A recent study revealed that MZF1-mediated MYC expression caused by wild-type LKB1 loss promotes tumor progression in lung adenocarcinoma cells [27]. Therefore, MZF1 has a dual functional property depending on the cellular environment. In our previous review article, we conclude that blocking MMP2 expression or activity may present a promising strategy for anticancer treatment in head and neck cancer [28]. Further studies are warranted to evaluate whether MZF1 repression of MMP2 transcription activity also exists in OSCC, and if so, the correlation between MZF1 and MMP2 could be clinically useful as a prognostic marker in OSCC diagnostics.

In this study, we performed immunohistochemical analysis and observed that negative nuclear expression of MZF1 was associated with advanced clinical stage and a larger tumor size in patients with OSCC. In addition, loss of MZF1 expression significantly related to poor overall survival according to the results of the Kaplan-Meier analysis. These results indicate that nuclear MZF1 is a potential biomarker for overall survival and OSCC progression.