Introduction

Oral cavity squamous cell carcinoma (SCC) may have a variable degree of aggression and propensity to metastasize to cervical lymph nodes. It is important to predict its malignant potential, as diffusion to cervical lymph nodes directly affects patients’ prognoses. The existing staging system of the American Joint Committee on Cancer seems to be insufficient for establishing the prognosis for oral carcinoma. Hence many studies have investigated additional predictive factors for metastasis development and carcinoma recurrence.

Further understanding of the molecular alteration in oral SCC will allow us to provide more accurate and useful prognostic markers. Since the discovery of the tumour suppressor p53, more than 15 different tumour suppressor genes have been identified [1]. Mammary serine protease inhibitor (MASPIN), a 42-kDa protein, is known to have tumour-suppressor functions. MASPIN belongs to the serine protease inhibitor (serpins) superfamily, which is categorized to inhibitory and non-inhibitory serpins [2]. MASPIN expression has been demonstrated in several normal human tissues, including breast, prostate, epidermidis, lung, and stromal cells of the cornea [3].

A number of findings support the inhibitory effects of MASPIN on human tumours. Cell surface-associated MASPIN is primarily responsible for MASPIN’s anti-invasive properties [3]. Several reports indicate that MASPIN can act as an inhibitor of angiogenesis by blocking it both in vitro and in vivo models [3, 4]. MASPIN has also been implicated in apoptosis, or programmed cell death [3, 5].

Although the original observations pointed to the association of reduced MASPIN expression with cancer progression, ensuing studies have revealed this correlation to be far more complex than originally concluded [2]. In spite of being regarded as a tumour suppressor gene, paradoxically, both a decrease and an increase of MASPIN levels have been described to parallel tumour progression [6]. The difficulties in interpreting this kind of information probably depend on the fact that MASPIN demonstrates a different localization pattern: it localizes primarily to the cytoplasm, but may also be located to the nucleus, the secretory vesicles and the cell surface. The protein subcellular partition appears to be important in its function [3]. Moreover, it has been noted that a nuclear localization, opposed to a combined nuclear-cytoplasmic localization of MASPIN, segregates with molecular markers associated with favourable outcomes, particularly in breast, ovarian and laryngeal cancer [6, 7]. It is reasonable to hypothesize that the apparently paradoxical expression pattern of MASPIN in tumour progression may be due to the regulation of its function at the level of subcellular localization.

The present study investigated the biological and prognostic role of MASPIN in relation to its subcellular localization in a series of consecutive cases of oral cavity SCCs.

Methods

Patients

A total of 56 consecutive cases of oral cavity SCC were retrospectively evaluated. Forty-six patients were male and ten female, with a mean age of 65.9 ± 12.6 years. In 17 patients primary carcinoma localization was mobile tongue, in 10 floor of the mouth (FOM), in 10 hard palate, in 7 retromolar area, in 7 cheek mucosa, in 5 mucosa of the lower lip. All patients underwent upper aero-digestive tract endoscopy and fibroscopic evaluation of the oesophagus. Neck ultrasonography (with or without fine needle aspiration cytology), head and neck contrast enhanced computerized tomography or/and magnetic resonance imaging, chest X-rays, liver ultrasonography, alkaline phosphatase determination were also performed.

All patients underwent surgical treatment of primary carcinoma and modified radical or selective neck dissection/s at the Section of Otolaryngology of the University of Padova in the period between 2002 and 2006. According to the TNM Classification of Malignant Tumours of the International Union Against Cancer [8], the pathologic staging of primary oral lesions (pT) was T1 in 24 cases, T2 in 20, T3 in 6, and T4 in 6. Pathological regional lymph node staging (pN) was N0 in 41 cases, N1 in 10, N2 in 4, and N3 in 1. Considering pathological grading, 17 out of 56 cases were staged as G1, 27 as G2, and 12 as G3. Post-operative radiotherapy was also performed in 21 cases with pathologically close surgical margins, evidence of perineural/lymphatic/vascular invasion, multiple positive nodes (three or more), pN3 staging, or extracapsular spread.

Immunohistochemistry

For immunohistochemical evaluation formalin-fixed, paraffin-embedded 5 μm sections were cut for all 56 cases.

The sections were pre-treated in a microwave oven (750 watt) for 20 min in a citrate buffer (10 mM, pH 6.0). For each sample, MASPIN (mouse monoclonal antibody, clone G167-70, BD-Biosciences, PharMingen International, San Diego, CA, 1:500 dilution) reactivity was evaluated by incubation. The sections were pre-incubated with protein block (Novocastra Laboratories Ltd, Newcastle upon Tyne, UK) for 5 min and stained with selected antibodies. Post-primary block (Novolink Polymer Detection System, Novocastra) was applied to the specimens for 20 min. Specimens were then washed with phosphate-buffered solution (PBS) (pH 7.0) for 3 min and incubated with Novolink Polymer for 20 min. The Automate Staining System (Biogenex-USA) was used. The colour was developed using 3,3′-diaminobenzidine (DAB) (DAKO, Glostrup, Denmark) for 4 min. The sections were counterstained with Meyer’s haematoxylin. We used human normal breast tissue as positive control for MASPIN. Negative and isotype controls at the same concentration were used in all staining runs.

The sections were scanned by the senior pathologist (S.B.) at 100× magnification in order to select the less differentiated areas of carcinoma, free of necrosis or haemorrhage. For MASPIN reactivity different evaluations were performed.

MASPIN subcellular pattern of distribution was scored as: negative, nuclear (when a distinctly prevalent nuclear reactivity was disclosed), cytoplasmic (when a cytoplasmic reactivity was distinctly prevalent), nuclear-cytoplasmic (when both cellular compartments were significantly interested). The percentage of stained nuclei (considering a minimum of 600 counted cells at 400× magnification) was determined. The presence of strong cytoplasmic MASPIN reactivity was investigated at 50× magnification. Weak cytoplasmic MASPIN reactivity was considered as negative. The percentage of positive cells was determined in cytoplasmic MASPIN-positive specimens (400× magnification).

Data analysis

When required, the following statistical tests and model were applied: Fisher exact test, Kruskal–Wallis test. The Kaplan–Meier survival function and the log–rank test were applied to display and evaluate the different disease-free intervals of the patients stratified according to MASPIN pattern of expression, nuclear MASPIN expression (%) and cytoplasmic MASPIN expression (presence/absence). A P value <0.05 was considered to be significant; values in the range 0.10 > P ≥ 0.05 were considered as indicating a statistical trend. The STATA 8 (StataCorp, College Station, TX, USA), statistical package, was used for all evaluations.

Results

Clinicopathologic features and prognosis (Table 1)

Out of 56 cases of oral cavity SCC, 37 did not experience malignancy recurrence after treatment. Nineteen patients developed loco-regional malignancy recurrence after a mean period of 15.6 ± 13.0 months. Mean follow-up time was 24.8 ± 16.4 months. Fisher exact test disclosed a significant association between pN-stage (pN0/pN+) and recurrence of disease (P = 0.032). The log–rank test found a significantly longer disease-free interval (in months) in pN0 patients than in pN+ ones (P = 0.038).

Table 1 Clinical and immunohistochemical evidences in considered series
Full size table

The Fisher exact test did not find significant association between carcinoma recurrence and pT-stage (P = 0.324), and pathological grading (P = 0.662). The log–rank test did not show a significant difference in the disease-free interval (in months) in patients stratified according to pT-stage (P = 0.179), or pathological grading (P = 0.865).

The group of patients who underwent post-operative irradiation was characterized by a higher incidence of disease recurrence after treatment (Fisher exact test, P = 0.001) and shorter disease-free interval (log–rank test, P = 0.005) than the non-irradiated group of patients.

MASPIN pattern of expression

Normal epithelial cells of oral cavity mucosa showed only weak cytoplasmic immunohistochemical staining for MASPIN (Fig. 1). Out of 56 oral cavity SCC, 23 specimens (41.1%) showed no MASPIN immunoreactivity (Fig. 2a). On the other hand, 3 out of 56 cases were characterized by a nuclear pattern of MASPIN compartmental localization (Fig. 2b), 1 case by a nuclear-cytoplasmic pattern (Fig. 2c), 29 cases by a cytoplasmic pattern (Fig. 2d). Statistical analysis ruled out significant association between MASPIN pattern of expression and pT-stage (Fisher exact test, P = 0.596), N-stage (Fisher exact test, P = 0.169), pathological grading (P = 0.149), or loco-regional recurrence of disease (Fisher exact test, P = 0.697). The log–rank test ruled out a significant difference in the disease-free intervals (in months) in patients stratified according to MASPIN pattern of expression in oral cavity SCC (P = 0.638).

Fig. 1
figure 1

Normal oral cavity mucosa near SCC showed weak cytoplasmic MASPIN expression

Full size image
Fig. 2
figure 2

a No MASPIN expression in oral cavity SCC specimen; b nuclear pattern of expression; c MASPIN nuclear-cytoplasmic pattern of expression; d MASPIN cytoplasmic pattern of expression

Full size image

MASPIN nuclear expression

The median MASPIN nuclear expression was 0% (mean value 3.9 ± 12.0%). Statistical analysis showed no significant association between MASPIN nuclear expression and pT-stage (Kruskal–Wallis test, P = 0.219), N-stage (Kruskal–Wallis test, P = 0.439), pathological grading (Kruskal–Wallis test, P = 0.120), or loco-regional recurrence of disease (Fisher exact test, P = 0.617). The log–rank found no significant association between disease-free intervals (in months) and MASPIN nuclear expression in oral cavity SCC (P = 0.491).

MASPIN cytoplasmic expression

Strong immunoreactivity for MASPIN in SCC carcinoma cells cytoplasm was found in 30 cases. In these cases, the percentage of positive cells for cytoplasmic MASPIN reactivity ranged from 10 to 100% (median 50.0%, mean 59.7 ± 29.4%). Out of 41 pN0 patients, 25 (61.0%) showed MASPIN cytoplasmic reactivity; 5 out of 15 pN+ patients (33.3%) showed MASPIN cytoplasmic reactivity. The Fisher exact test found a significant association between MASPIN cytoplasmic expression [presence (1)/absence (0)] and pN-stage (pN0/pN+) (P = 0.032). No significant associations were found between MASPIN cytoplasmic expression and pT-stage (Fisher exact test, P = 0.422), pathological grading (Fisher exact test, P = 0.473) or loco-regional recurrence of disease (Fisher exact test, P = 0.265). The log–rank found no significant association between disease-free intervals (in months) and MASPIN cytoplasmic expression in oral cavity SCC (P = 0.283).

Discussion

MASPIN shows a broad localization pattern: it may localize to the cytoplasm, but also to the nucleus, the secretory vesicles and the cell surface. Interestingly, in the earlier studies, the nuclear detection of MASPIN was thought to be only an artefact [6]. MASPIN function in the nucleus is less investigated and probably more complex; in the absence of a nuclear localization signal, MASPIN must either be chaperoned to the nucleus or cross the nuclear membrane by passive diffusion. Although the significance of the subcellular location of MASPIN is still not clear, the partitioning of MASPIN into different subcellular locations is indicative of different functions, thus distinct binding partners [2, 4].

A very limited number of studies tried to confirm that MASPIN biological functions are linked to protein’s subcellular localization in head and neck squamous cell carcinomas. Three studies of our group investigated the role played by MASPIN in laryngeal SCC and its prognostic value. Marioni et al. [7] found a significantly higher MASPIN nuclear expression in the group of patients without SCC recurrence after surgery than in the group with recurrence. Nuclear localization of MASPIN was significantly associated with a longer disease-free interval in the considered cohort of patients. Nuclear MASPIN expression was inversely correlated with Ki67 expression in laryngeal SCC cells. The above mentioned investigation was one of the first studies taking into account the sub-cellular localization of MASPIN and suggesting that the compartmental localization of MASPIN could be indicative of different cellular effects, with the localization in the nucleus being associated with less aggressive laryngeal SCC properties. The results of a recent investigation supported a crucial role of sub-cellular nuclear localization of MASPIN in laryngeal SCC angiogenesis [9]. Marioni et al. [5] studied for the first time the role of MASPIN in laryngeal carcinoma apoptotic mechanism investigating the relation between MASPIN expression and the apoptosis marker M30. The outcomes support the hypothesis of an important role of sub-cellular nuclear localization of MASPIN in its apoptosis-sensitizing effect in laryngeal SCC.

To the best of our knowledge, only four previous studies investigated the expression and role of MASPIN in oral SCC. The expression of MASPIN in 44 oral SCC patients’ specimens was detected by Xia et al. [10]. They found that 66% of the cases expressed low to intermediate levels of MASPIN and 34% of the cases expressed high levels of MASPIN. Cytoplasmic, but not membrane, staining was observed. High MASPIN expression was significantly associated with the absence of nodal metastasis. The patients with high MASPIN expression had better survival rates than those who had low or intermediate levels of MASPIN expression. Xia et al. [10] further examined by Western blot analysis MASPIN protein expression in a series of six head and neck SCC cell lines, and found that all but one expressed low or no MASPIN protein. Yasumatsu et al. [11] investigated MASPIN expression in 37 biopsy specimens of SCC of the tongue obtained from patients who underwent brachytherapy. MASPIN immunostaining was observed in the cytoplasm of tumour cells. No correlation was found between MASPIN expression and clinical stage, or SCC growth pattern. On the other hand, the absence of MASPIN expression was found more frequently in cases with subsequent cervical lymph node metastasis. The clinical follow-up data showed a significantly longer disease-free interval and overall survival periods for MASPIN-positive than for MASPIN-negative cases. Cho et al. [12] investigated the expression of MASPIN, mutant-type p53 and VEGF by immunohistochemistry in 33 patients with stages I and II oral tongue cancer. MASPIN displayed a cytoplasmic staining pattern. Of the 33 considered specimens, 10 showed high intensity and 23 low intensity staining. No correlation was found between MASPIN expression and recurrence rate. Survival curves were not different significantly between high and low-intensity groups of MASPIN immunostaining. Very recently, Iezzi et al. [13] immunohistochemically evaluated MASPIN expression in oral carcinoma of 89 patients. MASPIN expression was significantly associated with tumour differentiation grade and lymph node status.

In the present study we considered 56 consecutive cases of oral cavity SCC. To reduce the possibility of significant bias due to series heterogeneity in a retrospective setting, only surgical specimens (not biopsies) of carcinoma of a specific structure (oral cavity) consecutively treated in a defined period (2002–2006) by the same surgical staff were investigated. Statistical analysis found a significant association between pN-stage and recurrence of disease (P = 0.032) and a significantly longer disease-free interval in pN0 patients than in pN+ ones (P = 0.038). As we had done in our previous research program investigating MASPIN expression in laryngeal SCC, we focussed on the biological and prognostic role of MASPIN in relation to its subcellular localization. In the present series, 23 out of 56 oral cavity SCC specimens (41.1%) showed no MASPIN immunoreactivity. The complete loss of MASPIN expression observed in these oral cavity SCCs may be due to the methylation-mediated transcriptional silencing of the MASPIN promoter [14]. Twenty-nine cases (51.2%) were characterized by a cytoplasmic compartmental pattern of expression. Nuclear and nuclear-cytoplasmic MASPIN patterns of expression were found only in three (5.3%) and one (1.8%) cases, respectively. Interestingly, this evidence differed from laryngeal SCC: 12.1% of the cases were characterized by a nuclear pattern of MASPIN compartmental localization and 19.7% by a nuclear-cytoplasmic pattern in Marioni et al. [5] series. In the present study, the statistical analysis found no significant associations between MASPIN pattern of expression and loco-regional recurrence of disease or disease-free interval after treatment. Sixty-one per cent of the cases without neck lymph node metastasis (pN0) were strongly MASPIN-positive in the cytoplasm of primary tumour cells, 33% of the pN+ cases were MASPIN-positive in the cytoplasm. The Fisher exact test found a significant association between MASPIN cytoplasmic expression and pN-stage.

Conclusions

Our previous observations supporting the hypothesis that MASPIN nuclear location is a good prognostic factor in laryngeal SCC, probably due to an important role of nuclear MASPIN in apoptosis-sensitizing effect and in reducing microvascular density, were not confirmed in oral cavity SCC. In the present series of oral SCC, 51% of the cases were characterized by a cytoplasmic compartmental pattern of MASPIN and 41% by no MASPIN immunoreactivity. Considering the statistical evidence of a significant association between MASPIN cytoplasmic expression and pN-stage, it could be hypothesised that negative MASPIN immunoreactivity in oral carcinoma cells cytoplasm may be useful as a clinical marker to identify patients at risk of disease disseminating to neck lymph nodes: a large series confirmation may imply the necessity for elective neck dissection in clinically N0 patients with primary oral SCC negative for MASPIN cytoplasmic immunoreactivity.

At the molecular level, the functional consequences that change in the subcellular localization of MASPIN may have and how they achieve cell transformation are questions that remain to be addressed in oral SCC.