Increased Expression of Macrophage Migration Inhibitory Factor During Progression to Hypopharyngeal Squamous Cell Carcinoma

Materials and Methods

Patient characteristics. A total of 81 patients with HSCC who underwent surgery aimed at curative tumour resection were studied. The patient files were compiled retrospectively (January 1989 to December 2001) from records of the ENT Department at the Hôpital Claude Huriez (Lille, France). Twenty percent of stage IV hypopharyngeal patients presented with positive margins. Description of tumour status was based on the histopathological grade of tumour differentiation (criteria defined in (23)) and the TNM staging classification (24). Detailed information on patient age, gender, tumour histopathology, type of hypopharyngeal surgery, response to treatment at the primary tumour site, follow-up data up to the last patient contact, and disease status were available for 81 patients (Table I). All tissue specimens were from patients who did not undergo chemotherapy or radiotherapy prior to surgery. All stage IV hypopharyngeal SCC patients received additional post-operative radiotherapy. The HSCC patient cohort was thus nearly homogeneous in terms of histopathological and clinical criteria. Patients suffering from SCCs localised at other sites of the head and neck area were excluded from the study. This study was approved by the local Institutional Review Board.

Antibody preparation and quality controls. Human MIF was purified by affinity chromatography as described previously (25), followed by high-resolution preparative gel electrophoresis. The protein was rigorously checked for purity by two-dimensional gel electrophoresis and mass spectrometry, and the purified MIF was used as an antigen for raising polyclonal antibodies. Antibody titers in rabbit serum were regularly monitored using ELISA. Serum was fractionated for immunoglobulin G by chromatography on an Agarose Fast Flow resin bound with recombinant protein A (Upstate Biotechnology, Millipore, Schwalbach, Germany).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Immunohistochemical staining profile for MIF in TF_E (A), Low_D (B) and High_D (C) in areas surrounding hypopharyngeal CA (D). Magnification is ×320, apart from the insert in C, where it is ×640.

Western blotting. Tissue from hypopharyngeal TF_E and HSCC biopsies was homogenised in 2.5 vol Tris-sucrose buffer (25 mM Tris-HCl, pH 7.4, containing 250 mM sucrose) with 5 mM EDTA and protease inhibitors. Tissue homogenates were centrifuged at 700×g for 10 min, then NaCl and MgSO₄ were added to the supernatants to reach final concentrations of 100 mM and 1 mM, respectively. These supernatants were spun at 100,000×g for one hour. Homogenisation and centrifugation were carried out at 4°C. Supernatants from the ultracentrifugation step were used as cytosolic fractions and assayed for protein content. For Western blot analysis, 20 μg of cytosolic proteins were resolved by SDS-PAGE on 8% T acrylamide-bisacrylamide gels. After separation, proteins were electrotransferred from the gels onto nitrocellulose membranes (Hybond ECL; Amersham Pharmacia Biotech Europe GmbH, Freiburg, Germany). Non-specific protein-binding sites on the membranes were blocked for three hours at room temperature using a blocking buffer (blotto A) [TBS buffer (10 mM Tris-HCl, pH 8, 150 mM NaCl) containing 5% non-fat milk and 0.05% Tween-20]. Membranes were then incubated overnight at 4°C with the anti-MIF antibody (2 μg/ml). Exposure to the anti-MIF antibody was followed by incubation for 2 hours at room temperature with a peroxidase-conjugated goat anti-rabbit secondary. Finally, after 15 seconds of incubation in the presence of BM chemiluminescence blotting substrate (POD), immunoreactive bands were visualised following exposure of the membrane to a sensitive film (Amersham Hyperfilm ECL, GE Healthcare Limited, Buckinghamshire, UK). Biotinylated molecular weight markers were run and blotted in parallel for internal calibration.

Immunohistochemistry. All tumour samples were fixed for 24 h in 10% buffered formaldehyde, dehydrated and embedded in paraffin. Immunohistochemistry was performed on 5 μm-thick sections mounted on silane-coated glass slides, as previously detailed (26). Before starting the immunohistochemistry protocol, dewaxed tissue sections were briefly subjected to microwave pretreatment in a 0.01 M citrate buffer (pH 6.0) for 2×5 min at 900 W. The sections were then incubated with a solution of 0.4 % hydrogen peroxide for 5 min to block endogenous peroxidase activity, rinsed in phosphate-buffered saline (PBS; 0.04 M Na₂HPO₄, 0.01 M KH₂PO₄ and 0.12 M NaCl, pH 7.4) and successively exposed for 20 min to solutions containing avidin (0.1 mg/ml in PBS) and biotin (0.1 mg/ml in PBS) to avoid false-positive staining reactions resulting from endogenous biotin. After thorough washing with PBS, the sections were incubated for 20 min with a solution of 0.5 % casein in PBS and sequentially exposed at room temperature to solutions of: (i) the specific primary anti-MIF antibody, (ii) the corresponding biotinylated secondary antibody (polyclonal goat anti-rabbit IgG) and (iii) the avidin-biotin-peroxidase complex (ABC kit). Between incubation steps, the samples were thoroughly washed to remove unbound proteins. Presence of antigen (MIF) in the sections was visualised by incubation with the chromogenic substrates containing diaminobenzidine and H₂O₂. After rinsing, the sections were counterstained with luxol fast blue and mounted. To exclude antigen-independent staining, the incubation step with primary antibody was omitted from the controls. In all cases, these controls were negative. The biotinylated secondary antibodies and ABC kit were obtained from DakoCytomation (Glostrup, Denmark).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Quantitative determination (with computer-assisted microscopy) of A: immunohistochemical MIF staining intensity (mean optical density and B: the percentage of tissue area staining for MIF (labelling index) in a series of 15 TF_E (tumour-free epithelium) samples, 29 low-grade dysplasia (Low_D) samples, 25 high-grade dysplasia (High_D) samples and 81 stage IV hypopharyngeal carcinoma (CA) samples.

Definition of low- and high-grade epithelial dysplasias. Morphological characteristics of dysplasia include increased cellular density associated with a large number of mitotic figures in the vicinity of the basal layer, irregular maturation, loss of polarity and dyskeratosis. Cytological dysplasia is characterised by an increased ratio of nuclear to cytoplasmic area, anisocytosis, poikilocytosis, nuclear polymorphism, chromatin condensation and large nucleoli; features sometimes associated with atypical mitotic figures. Low-grade dysplasia, consisting of mild and moderate dysplasia, presents atypical features extending over the lower or middle third of the epithelium. High-grade dysplasia and carcinoma in situ extend over the entire thickness of the epithelium (27).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Expression of MIF in well- (A) and poorly differentiated (B) hypopharyngeal HSCC tissue samples and mean labelling index values of the tissues; (C) Error bars denote standard errors.

Computer-assisted microscopy. Following the immunohistochemical steps, quantitative features of MIF staining were determined using a computer-assisted KS 400 imaging system (Carl Zeiss Vision, Hallbergmoos, Germany) connected to a Zeiss Axioplan microscope, as detailed previously (26). For each microscopic field, the analysis was focused on neoplastic cells or tumour-free epithelia using computer-assisted morphometry after interactive identification. These tissue areas were delimited precisely with the computer mouse. In each case, 15 fields were scanned covering a surface area ranging from 60,000 to 120,000 μm². The quantitative analysis of immunohistochemical staining yielded data on the following two variables: (i) the labelling index (LI), defined as the percentage of positive cells, and (ii) the mean optical density (MOD), defined as the staining intensity of positive cells (26). For each type of dysplasia (low– or high–grade), the respective fields within the peritumoural areas were defined by one of the authors (XL) with expertise in this diagnostic procedure.

Data analysis. Independent groups of quantitative data were compared using the non-parametric Kruskall-Wallis (more than two groups) or Mann-Whitney U tests (two groups). In the case of more than two groups, post-hoc tests (Dunn procedure) were used to compare pairs of groups (to avoid multiple comparison effects). Relationships between the qualitative (or ordinal) variables analysed were studied by contingency tables. The level of significance of correlations was evaluated by the χ² or the exact Fisher test (in the 2×2 cases). A decision-tree approach was systematically applied to disclose threshold values aiming at separation of patient groups with and without recurrence (28). Statistical analyses were carried out using Statistica software (Statsoft, Tulsa, USA).