Cell cycle association and hypoxia regulation of excision repair cross complementation group 1 protein (ERCC1) in tumor cells of head and neck cancer

The procedures followed were in accordance with the ethical standards of the committee on human experimentation of the institution or in accord with the Helsinki Declaration of 1975 as revised in 1983. Permission was obtained from the local ethics committee to collect pretreatment biopsy samples for paraffin embedding, sectioning, and immunohistochemical analysis (Reference Number: UN4428 303/4.14). Patients with locally advanced unresectable squamous cell carcinomas of the head and neck region treated between April 2008 and October 2013 with primary RCT at the Department of Otorhinolaryngology—Head and Neck Surgery, Medical University Innsbruck, were consecutively included. Pretreatment tumor samples were obtained during diagnostic panendoscopy. Tumor biopsies were immunohistochemically stained with the ERCC1 antibody 8 F1 from Acris (Herford, Germany). The proportion of ERCC1⁺ cells and nuclei in tumor cell nests was counted in 10 high power fields. All patients received primary RCT. Treatment response was assessed 6–8 weeks after the end of treatment and grouped into either complete response or treatment failure, because incomplete responses were associated with poor survival. Clinical data are summarized in Supplementary Table 1. Twenty-six normal mucosa tissue specimens were derived from uvulopalatopharyngoplasties (UPPP) and other nontumor samples.

Patients underwent treatment upon multidisciplinary tumor board recommendation. The standard treatment for advanced HNSCC not suited for surgical treatment was primary radiochemotherapy (pRCT). Standard pRCT consisted of percutaneous normofractionated computer tomography CT-supported 3D-planned radiotherapy (RT) with 70–72 Gy, combined with CDDP or mitomycin C and 5-fluorouracil according to common protocols [17‐19]. In patients not eligible for chemotherapy, RT with concomitant weekly cetuximab or—in case of cetuximab intolerance—a hyperfractionated RT protocol was administered [20, 21] (detailed in Supplementary Table 1). A complete diagnostic workup including biopsies from the initial tumor regions was performed 8 to 12 weeks following end of treatment. Moreover, normal mucosa samples of 26 nontumor patients were included. With regard to the study aim, these patients had no relevant mucosal abnormalities.

Five-micrometer-thick paraffin sections were dewaxed and antigen retrieval was performed in a Discovery^TM automated staining system (Ventana, Tucson, AZ, USA). ERCC1 antibodies were added to the sections by manual titration (mouse monoclonal clone 8 F1 Acris, 1:100 dilution), and the staining was completed by the Discovery^TM automated staining system using universal secondary antibody solution, hematoxylin counterstaining, and the DAB MAP Kit (all Ventana products). All sections were stained with control mouse and rabbit immunoglobulins, using the same highest concentration as for the primary antibodies, and these controls were not reactive [22]. The immunohistochemical reactions were observed independently by two blinded observers, who collected 10 representative tumor cell nests from each specimen [22], or 10 random areas of normal mucosa (UPPP samples). These regions were analyzed on an Olympus BX50 microscope (Tokyo, Japan) and on the TissueFaxs system (Tissuegnostic, Vienna, Austria). The whole number of cells and the number of immunopositive cells were counted in 10 high power fields (HPF). The staining index (percent of positive cells) was calculated as the average ratio between the positive cells and whole cell number [22]. Correspondingly, the nuclear staining index was calculated for cell nuclei only.

Paraffin sections of 18 HNSCC specimens (Supplementary Table 2) were used to compare the immunohistochemical reactivity of two ERCC1 antibodies: mouse monoclonal, clone 8 F1 from Acris (1:100 dilution), and rabbit polyclonal, FL-297 from Santa Cruz Biotechnology (Santa Cruz, CA, USA, 1:250 dilution) as previously described and suggested by Niedernhofer and coworkers [23‐25]. For immunohistochemistry of XPF, the mouse monoclonal antibody SPM228 (1:1,000; Abcam, Cambridge, MA) was used [26].

BEAS-2B immortalized noncancer bronchial epithelial cells [27] and Detroit 562 metastatic pharyngeal squamous cell carcinoma cells were used [28]. Both cell lines were of commercial origin, and their culture conditions have been published before [28]. In some western blot experiments, SCC-25 oral squamous cell carcinoma cells were used.

Hypoxic cell culture conditions were performed in a Memmert INCO 108 incubator (Schwabach, Germany), which provides controlled 1 % O₂. Detroit 562 cells were plated at 5 × 10⁴/ml in six-well plates at 3 ml/well. After an initial 48-h culture period, one plate was used as “control zero” for RNA isolation and real-time PCR. The further plates were used for RNA isolation and protein fractionation in three repeats. Cells were cultured after the initial 48-h normoxic culture for 48–96 h normoxic, or for 48–72 h in hypoxic conditions, and also combined, 48 h hypoxic followed by 24 or 48 h normoxic conditions. These culture conditions were based on a published reference [29]. Gene expression of ERCC1 was analyzed related to the “control zero” condition.

For protein synthesis analysis, cells were plated at 5 × 10⁴/ml in six-well plates (3 ml cell suspension/well), grown for 4 days, and on the fifth day, cells in one well were scraped into 500 μl extraction buffer followed by nonnuclear and 0.35 M NaCl nuclear protein fractionation as described previously [30]. Also, cells in hypoxic or normoxic or combined culture were used for nuclear protein fractionation. Nuclear and nonnuclear protein fractions were subjected to protein concentration measurement, and 10 μg proteins were used for western blots with antibodies of SPM228 (XPF, in 1:1,000 dilution), 8 F1 (ERCC1, mouse monoclonal, in 1:250 dilution), and FL-297 (ERCC1, rabbit polyclonal, in 1:200 dilution). Blots were processed in a western blot protocol previously detailed [31].

BEAS-2B and Detroit 562 cells were plated at 5 × 10⁴/ml in 96-well plates, 100 μl/well, were cultured for 48 h, and were treated with 0.1–100 μmol/l CDDP [32] or MMC [33] or 5-FU [34] for 48 h in complete medium (with medium replacement after 24 h), according to previous publications [33‐35]. After 48 h of treatment, the medium was replaced with a fresh one, and after 24 h, the (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbromide (MTT) assay was performed as described previously [35]. The formazan crystals were dissolved in 10 % sodium dodecyl sulfate- and 10 mM HCl-containing solution, and the absorption was measured at 550 nm using an Anthos 2010 ELISA reader (Salzburg, Austria). The absorption values of cell-free blank wells were removed from the absorption values of cell-containing wells, and the average absorption value of untreated cells was considered as 100 % growth capacity. The absorption values of the treated cells were related to the controls as percentage of growth capacity. The relationship between growth capacity and treatment drug concentrations was analyzed by regression using nonlinear dose-response fitting methods of GraphPad Prism 4.03 (GraphPad Software Inc, San Diego, CA, USA), and concentrations were determined, which cause 50 % growth inhibition compared to untreated cells (IC-50).

ERCC1 gene expression at the messenger RNA (mRNA) level has been determined after treatment with IC-50 doses of CDDP, MMC, and 5-FU. BEAS-2B and Detroit 562 cells were plated at 5 × 10⁴/ml in 10 cm cell culture dishes, 10 ml/dish, and were then cultured and treated as described by the IC-50 determination with IC-50 concentrations of CDDP or MMC or 5-FU for 48 h in complete medium (with medium replacement after 24 h). After 48 h of treatment, the medium was replaced with a fresh one, and after 24 h of culture in complete medium, the cells were used for immunocytometry [36] or for RNA isolation [37‐39]. For RNA isolation, the same cells and treatments were performed as for immunocytometry. RNA was reverse-transcribed as published before [31, 38], and gene expression levels of ERCC1 were determined according to previous reports [31, 38‐40]. All used human PCR primers were published previously: ERCC1 [41]; β-actin [42], which was used as a housekeeping gene, was not regulated in control and treated samples and also not in hypoxic cultures. The relative gene expression was calculated as previously reported [42].

After the above-detailed cell culture and treatments, cells were trypsinized and suspended in 750 μl ice-cold phosphate buffered saline (PBS)/MgCl₂ (PBS containing 2 mM MgCl₂) / 10⁶ cells, and 250 μl cold 1 % BD-Cellfix (Becton Dickinson, Vienna, Austria) was added under continuous vortexing [36]. This mixture was incubated on ice for 1 h, and the cells were centrifuged at 1,200 rpm for 5 min at 4 °C, followed by 15 min incubation in 0.2 % Tween-20/PBS/10⁶ cells at 37 °C. The cells were centrifuged at 1,200 rpm for 5 min at 4 °C again and then incubated in 5 % bovine serum albumin in PBS/MgCl₂ at room temperature for 20 min followed by centrifugation at 1,200 rpm for 5 min at 4 °C. The cells were resuspended in 1 % bovine serum albumin in PBS/MgCl₂ and were incubated either with isotype-specific control mouse IgG2b, or with ERCC1 (8 F1, Acris) antibody at 100 times dilution at room temperature for 90 min followed by centrifugation at 1,200 rpm for 5 min at 4 °C. The immunocytochemical reaction was detected by 100 times diluted anti-mouse Alexa 488 (Jackson ImmunoResearch, Newmarket, Suffolk, UK) incubated in 1 % bovine serum albumin in PBS/MgCl₂ at room temperature for 30 min. Subsequently, DNA was stained with propidium iodide for cell cycle analysis as published before [35]. The double-stained cells were flow cytometrically examined on a Coulter Epics XL-MCL (Beckman Coulter, Brea, CA, USA). In addition to the cells that reacted through the whole procedure, the following controls were used for setting up the flow cytometry conditions and for compensation: cells reacted with whole antibodies but without propidium iodide and anti-mouse Alexa 488, cells immunoreacted with isotype-specific control mouse IgG2b and DNA-stained with propidium iodide, and cells reacted with ERCC1 (8 F1, Acris) antibody but DNA not stained with propidium iodide. Frequency diagrams (histograms; fluorescence intensity on the X-axis, events on the Y-axis, or fluorescence intensity on the X-axis, events related to all events on the Y-axis) and scattergrams were recorded for Alexa 488 at the fluorescence-1 (FL-1), and for propidium iodide at the fluorescence-1 (FL-3) channel of a Beckman Coulter XL-MCL. Doublets and adhered cells were removed from the DNA analysis using FL3 signal height/signal area graphs.

Each experiment was performed in four independent sets containing four biological repeats/set. The raw experimental results were tested for normal distribution by D’Agostino and Pearson omnibus normality test using the GraphPad Prism 4.03 (GraphPad Software Inc, San Diego, CA, USA). Average values were compared by nonparametric tests (Mann-Whitney) or by Student’s t tests depending on the distribution of the data. Logistic analysis was performed using MedCalc 12.4 (Ostend, Belgium). Correlation analysis of the staining index of ERCC1 and XPF antibodies was performed by the “Pearson r” method using GraphPad Prism 4.03.