Small interfering RNA targeting epidermal growth factor receptor enhances chemosensitivity to cisplatin, 5‐fluorouracil and docetaxel in head and neck squamous cell carcinoma

Abstract

Overexpression of epidermal growth factor receptor (EGFR) has been found in various epithelial malignancies, including head and neck squamous cell carcinoma (HNSCC), and is associated with increased tumor growth, metastasis, resistance to chemotherapeutic agents and poor prognosis. As such, EGFR is a potential target for antitumor therapy and several EGFR inhibitors have been investigated in preclinical or clinical settings. In the present study, we used small interfering RNA (siRNA) to downregulate EGFR expression while evaluating the effect of EGFR siRNA on cell proliferation, and the combined effects with cisplatin, 5‐fluorouracil (5‐FU) and docetaxel in HNSCC. Furthermore, HNSCC xenografts were treated with EGFR siRNA alone or in combination with cisplatin, and tumor growth was examined. EGFR expression, proliferation, angiogenesis and apoptosis index were evaluated by immunohistochemistry. The results showed that EGFR siRNA efficiently downregulated EGFR expression and inhibited cell growth of HNSCC. Treatment with EGFR siRNA in combination with cisplatin, 5‐FU and docetaxel enhanced chemosensitivity with a significant increase in apoptosis. EGFR siRNA delivered by atelocollagen enhanced the antitumor effect of cisplatin in the HNSCC xenograft model. These cumulative results suggest that EGFR siRNA combined with cisplatin, 5‐FU and docetaxel may be a feasible strategy to enhance the effects of chemotherapy in patients with HNSCC. (Cancer Sci 2006; 97: 1115–1124)

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer in the world and a major cause of morbidity.^{( 1} , ^{2 )} Treatment of HNSCC has primarily relied on classical modalities including surgery, radiation and chemotherapy or a combination of these methods, but the outcome has not improved significantly. Therefore, new treatment strategies are awaited, especially regarding gene therapy, with several gene therapy strategies for HNSCC currently under investigation in preclinical and clinical settings.^{( 3} , ^{4 )}

Epidermal growth factor receptor (EGFR) is a member of the ErbB family of receptors and is abnormally activated and commonly overexpressed in a number of human solid tumors including brain, breast, colon, lung, prostate, ovarian and head and neck cancers.^{( 5} , ⁶ , ^{7 )} Overexpression of EGFR has been reported in more than 80% of HNSCC tumors,^{( 8} , ^{9 )} and is associated with increased tumor growth, metastasis, resistance to chemotherapeutic agents and poor prognosis.^{( 10} , ¹¹ , ¹² , ^{13 )} Therefore, a variety of strategies to block the function of EGFR have been developed to inhibit tumor growth and thereby improve the overall clinical outcome. These include downregulation of EGFR expression, blockade of EGFR downstream signaling and abrogation of EGFR activation, using antisense oligonucleotides, monoclonal antibodies and tyrosine kinase inhibitors.^{( 4} , ^{14 )} Although promising results have been observed in preclinical studies using EGFR inhibitors, their efficiency is still far from satisfactory when used as monotherapy for cancer patients.^{( 15} , ^{16 )}

In recent years, RNA interference technology using small interfering RNA (siRNA) has rapidly become a highly specific and powerful tool as a method of target gene silencing.^{( 17} , ^{18 )} To date, a number of studies have demonstrated that siRNA can effectively and specifically downregulate oncogene expression.^{( 19} , ²⁰ , ^{21 )} Several studies comparing siRNA with antisense oligonucleotides have indicated that siRNA has advantages over antisense oligonucleotides in inhibiting gene transcription and resistance to nucleases.^{( 22} , ^{23 )} These results suggest that siRNA has the potential to be used as a targeted therapeutic agent in the treatment of cancer. In the present study, we investigated whether the downregulation of EGFR expression by siRNA inhibits tumor cell growth and enhances the chemotherapeutic effects of cisplatin, 5‐fluorouracil (5‐FU) and docetaxel, the chemotherapeutic agents used currently in the treatment of patients with HNSCC.

Materials and Methods

Cell lines and reagents.  Human head and neck squamous carcinoma cell lines HSC‐2 (JCRB0622) and SAS (JCRB0260) were purchased from the Human Science Research Resources Bank (Osaka, Japan). These cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen, Carlsbad, CA, USA) containing 10% heat‐inactivated fetal bovine serum (FBS; Sigma‐Aldrich, St Louis, MO, USA). Cisplatin, 5‐FU and docetaxel were obtained from Bristol‐Myers (Tokyo, Japan), Kyowa Hakko Kogyo (Tokyo, Japan) and Aventis Pharma (Tokyo, Japan), respectively.

siRNA synthesis.  Small interfering RNA duplexes were designed to target sequences on human EGFR mRNA (accession no. X00588) corresponding to nucleotides (nt) 2307–2325 (5′‐GAAGGAAACUGAAUUCAAA‐3′). The selected sequence was screened in a BLAST search against all of the known human genes to verify that only human EGFR mRNA was targeted. siRNA duplexes were obtained from Dharmacon (Lafayette, CO, USA). Each RNA contained two additional nucleotide 3′‐UU overhangs on the sense and antisense strands, and 5′‐phosphate on the antisense strand. In addition, scrambled EGFR siRNA was used as a negative control and obtained from Takara Bio (Shiga, Japan). The sequences of the oligonucleotides used were as follows: EGFR sense strand, 5′‐GAAGGAAACUGAAUUCAAAUU‐3′; EGFR antisense strand, 5′‐PUUUGAAUUCAGUUUCCUUCUU‐3′; control sense strand, 5′‐AGUAAUACAACGGUAAAGAUU‐3′, control antisense strand, 5′‐PUCUUUACCGUUGUAUUACUUU‐3′.

siRNA transfection in vitro.  Head and neck squamous cell carcinoma cell lines were plated in a 24‐well plate in 500 µL DMEM at a density of 2 × 10⁴ cells per well. siRNA duplexes were complexed with the TransIT‐TKO Reagent (Mirus, Madison, WI, USA) in Opti‐MEM I (Invitrogen) serum‐free medium, as described by the manufacturer's protocol. In brief, 2 µL of TransIT‐TKO was added to 50 µL Opti‐MEM I and allowed to sit at room temperature for 12 min. siRNA duplexes were added to the diluted TransIT‐TKO reagent and were incubated for 15 min at room temperature. Aliquots were added to the 24‐well plate containing 250 µL of DMEM. The final siRNA concentration was 50 nM in a 24‐well plate.

Flow cytometric analysis of EGFR cell surface expression.  For the detection of EGFR expression by flow cytometry, the cells were detached using trypsin (0.25%)–ethylenediaminetetracetic (0.02%) acid, washed once with medium, and then resuspended in phosphate‐buffered saline (PBS) with 2% FBS and 0.05% sodium azide (FACS buffer). Aliquots of 2 × 10⁵ cells were incubated at 4°C for 30 min with 5 mg/mL primary antibody B4G7 (mouse antihuman monoclonal anti‐EGFR antibody; supplied by Dr N. Shimizu, Keio University School of Medicine, Tokyo, Japan), herceptin (humanized anti‐HER2 monoclonal antibody; Chugai Pharmaceutical Co., Tokyo, Japan) or IgG control antibody (Sigma‐Aldrich). The cells were then washed and stained with phycoerythrin (PE)‐conjugated antimouse IgG (Sigma‐Aldrich) or fluorescein‐isothiocyanate (FITC)‐conjugated antihuman IgG (Sigma‐Aldrich). The expression of surface molecules was analyzed by FACS Caliber (Becton Dickinson Immunocytometry System, Mountain View, CA, USA) and the results were represented as the geometric mean fluorescence.

Cell proliferation assay.  To determine the effect of EGFR siRNA on cell proliferation, HSC‐2 and SAS cells transfected with control or EGFR siRNA (50 nM) were seeded in 12‐well plates at a density of 2 × 10⁴ viable cells per well. Mock transfected cells were treated with TransIT‐TKO, but no siRNA. At the indicated time points, the cells were trypsinized and counted using a hemocytometer using the trypan blue dye exclusion method in triplicate samples. To determine the growth inhibitory effects of EGFR siRNA in combination with cisplatin, 5‐FU or docetaxel, siRNA‐transfected (50 nM) cells and mock transfected cells were incubated for an additional 48 h after adding these chemotherapeutic agents, and then the viable cells were counted.

Cytotoxicity assay.  In vitro cytotoxicity was assessed by the 3‐(4,5‐dimethylthiazol‐2‐yl)‐5‐(3‐carboxymethoxyphenyl)‐2‐(4‐sulfenyl)‐2H‐tetrazolium, inner salt (MTS) assay using the CellTiter 96 Aqueous One Solution Proliferation Assay (Promega, Madison, WI, USA), according to the manufacturer's protocol. In brief, the cells transfected with either the indicated siRNA (50 nM) or mock transfected control were seeded in 100 µL of medium/well in 96‐well plates at a density of 2–3 × 10³ cells per well. After overnight incubation, the investigational agents were added at various concentrations. After treatment for 48 h, 20 µL of MTS was added to each well and the plates were incubated for 2 h at 37°C. The absorbance of each well was measured using an automated plate reader and the resulting optical density (OD) was proportional to the number of viable cells. Values for control cells were considered as 100% viability. The growth inhibitory effects of each agent alone and their combination were analyzed by generating concentration effect curves as a plot of fraction of unaffected cells versus drug concentration. The dose–response curves were plotted as a percentage of the control cell absorbance. The 50% inhibitory drug concentration (IC₅₀) was defined as a 50% reduction in absorbance in comparison to no‐treatment controls. The fold enhancement in chemosensitivity was considered to be the ratio of the IC₅₀ of the untransfected cells to the indicated siRNA oligonucleotide‐transfected cells. All of the experiments were carried out at least three times in triplicate.

Flow cytometric analysis of apoptosis.  DNA fragmentation was measured using flow cytometric analysis. In combination treatments, the cells transfected with either control or EGFR siRNA were exposed to cisplatin, 5‐FU or docetaxel (IC₅₀) for 48 h. Next, the cells were stained using the CycleTEST DNA Reagent Kit (Becton Dickinson Immunocytometry System), according to the manufacturer's protocol. In brief, the cells were pelleted, washed with ice‐cold PBS and then gently resuspended in 40 µL of citrate buffer (250 mM sucrose, 40 mM trisodium citrate dihydrate). Solution A (180 µL; trypsin in a spermine tetrahydrochloride detergent buffer) was added to 20 µL of cell suspension and the test tubes were inverted. After a 10‐min incubation at room temperature, 150 µL of solution B (ribonuclease A and trypsin inhibitor) was added to each tube with gentle mixing. After a second 10‐min incubation period, 150 µL of solution C (propidium iodide and spermine tetrahydrochloride in citrate stabilizing buffer) was added. The cells were incubated for 10 min on ice and then analyzed for DNA histograms by flow cytometry. The data were analyzed using the CELL Quest computer program (Becton Dickinson Immunocytometry System). Cells with fractional DNA content located on DNA frequency histograms to the left of the G₁ peak (sub‐G₁ cells) were identified as apoptotic cells.

Xenograft models in nude mice.  SAS cells (3 × 10⁶) were suspended in 100 µL of PBS and injected subcutaneously (s.c.) into the right flank of 4‐week‐old female BALB/c nu/nu mice (Clea Japan, Tokyo, Japan). After 1 week, when the tumors reached a volume of approximately 100 mm³, mice were divided randomly into multiple treatment groups as follows: no treatment; atelocollagen alone; cisplatin alone; EGFR or control siRNA with atelocollagen; and EGFR or control siRNA with atelocollagen plus cisplatin treatment groups. For the in vivo transfection of siRNA, we used atelocollagen (Koken Co., Tokyo, Japan) which was reported by Ochiya et al.^{( 24 )} For preparing the siRNA–atelocollagen complex, atelocollagen (0.5% in PBS at pH 7.4) and siRNA solution were mixed by rotation for 6 h at 4°C. The final concentration of atelocollagen was 0.375%. The intratumoral injection of siRNA (5 µM) mixed with atelocollagen (50 µL) was administered every 4 days after the first injection. Cisplatin (1 µg/g) was administered via intraperitoneal (i.p.) injection after 48 h of siRNA treatment. The animals were treated until the tumors ulcerated or reached a maximum diameter of 2 cm. At regular intervals, the tumor diameters were measured with a caliper and tumor volumes were calculated using the formula:

length × width² × 0.5.

All animals were treated in a humane manner and managed according to the regulation of our animal facility.

Immunohistochemical analysis.  For an immunohistochemical evaluation of EGFR, Ki‐67 and CD34, serial 4‐µm cryostat sections were obtained from representative paraffin blocks. Formalin‐fixed and paraffin‐embedded tissue sections were deparaffinized in xylene, rehydrated in a graded alcohol series and transferred to PBS. EGFR immunostaining was carried out using the EGFR pharmDx IHC kit system (Dako Japan, Kyoto, Japan), according to the manufacturer's instructions. For Ki‐67 and CD34 immunostaining, sections were incubated with 2% hydrogen peroxidase in methanol for 45 min to block endogenous peroxidase. Following PBS rinses, the sections were immersed in 10 mM citrate buffer (pH 6.0) and heated in a microwave for 25 min. The microwave‐irradiated sections were cooled to room temperature and washed with distilled water. Next, the sections were incubated in protein blocking solution (5% normal serum, FBS, in PBS) for 30 min. The sections were incubated for 2 h at room temperature with the primary antibodies mouse monoclonal anti‐Ki67 (clone MIB‐1; Dako Japan) diluted 1 : 100 in PBS and mouse monoclonal anti‐CD34 (clone MEC14.7; Hycult Biotechnology, Uden, the Netherlands) diluted 1 : 10 in PBS. No primary antibody was added to the sections used as negative controls. After PBS rinses, the sections were treated with horseradish peroxidase‐conjugated secondary antibody, rabbic anti‐mouse IgG, (Chemicon International, Temecula, CA, USA) diluted 1 : 500 in PBS for 1.5 h at room temperature. The reaction was developed with diaminobenzidine (Dako Japan) and then the slides were counterstained with Mayer's hematoxylin. Control samples exposed to secondary antibody alone showed no specific staining. For an evaluation of the proliferation index (PI) in each section, Ki‐67‐stained cells were quantified in at least five randomly selected fields at ×400 magnification. The fraction (%) of the cells that showed positive nuclear staining for Ki‐67 antigen was considered to be the PI. Microvessel density was estimated as the mean number of vessels stained with CD34 in at least five randomly selected microscopic fields at ×400 magnification. All positively stained discrete cells or cell clusters with or without visible lumens were counted as one microvessel. Determination of apoptosis by terminal deoxynucleotidyl transferase‐mediated deoxyuridine triphosphate brotin nick‐end labeling (TUNEL) staining was carried out using the ApopDETEK Cell Death labeling and Detection System (Enzo Life Sciences, Farmingdale, NY, USA), according to the manufacturer's instructions. Briefly, nuclear proteins were stripped from DNA by incubation in proteinase K for 10 min, and endogenous peroxidase was blocked with 3% H₂O₂ for 10 min. The sections were incubated in equilibration buffer for 10 min at room temperature followed by incubation in a buffer containing terminal deoxynucleotidyl transferase and biotin‐labeled dUTP for 1 h at room temperature. After PBS rinses, the sections were incubated with antibiotin‐conjugated peroxidase for 30 min. Brown color was developed with incubation in diaminobenzidine solution for 10 min at room temperature. For evaluation of the apoptotic index in each section, TUNEL‐stained cells were quantified in at least five randomly selected fields at ×400 magnification and the apoptotic index was calculated as the number of positive cells/total number of cells × 100.

Statistical analysis.  A statistical analyses were carried out using Prism 4 (GraphPad Software, San Diego, CA, USA) to analyze the data and to determine the IC₅₀. Statistical significance was determined using the Mann–Whitney test or unpaired t‐test. P < 0.05 was considered to be statistically significant.

Results

EGFR siRNA can efficiently and specifically suppress the level of EGFR expression in HSC‐2 and SAS cells.  We first evaluated the cell surface EGFR expression in HSC‐2 and SAS cells by flow cytometric analysis. A431 cells, which are known to overexpress EGFR, were used as a positive control. Similar EGFR expression levels were detected in all three cell lines. These results indicated that EGFR protein was highly expressed at the cell surface of both HSC‐2 and SAS cells (Fig. 1a). Subsequently, we examined the effects of EGFR siRNA on the expression of EGFR. Cells were transfected with annealed 21‐mer sense and antisense RNA oligonucleotides (50 nM) directed against EGFR or its scrambled siRNA. The ability of EGFR siRNA to suppress EGFR expression was confirmed by flow cytometry. A flow cytometric analysis revealed the expression of EGFR to be potently suppressed by EGFR siRNA, but not by scrambled EGFR siRNA (Fig. 1b,c), and EGFR siRNA did not affect the expression of HER2 (Fig. 1d) in either HSC‐2 or SAS cells. These results indicated that EGFR siRNA could efficiently and specifically suppress EGFR expression in HSC‐2 and SAS cells.

Figure 1.

Effects of epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) on the expression of EGFR in HSC‐2 and SAS cells. (a) EGFR is overexpressed in HSC‐2 and SAS cells. Cell surface EGFR expression in HSC‐2 and SAS cells was determined by flow cytometry after staining with phycoerythrin (PE)‐conjugated antihuman EGFR monoclonal antibody (B4G7). Filled histograms represented non‐specific staining with mouse Ig‐PE (negative control). (b) Scrambled control siRNA did not significantly affect the expression of EGFR compared to mock transfected cells, whereas EGFR siRNA induced marked downregulation of EGFR expression. Bold line histograms represent the proportion of cells that were transfected with control (upper panel) or EGFR siRNA (lower panel). Filled histograms represent the proportion of mock transfected cells that were treated with TransIT‐TKO, but no siRNA (positive control). (c) EGFR expression was quantified by flow cytometry. The results were expressed as the percentage of intensity of fluorescence relative to controls (*P < 0.05). (d) HER2 expression was unaffected by either siRNA treatment. Filled histograms represented non‐specific staining with the secondary antibody without prior staining with the primary antibody (negative control). Thin line histograms represent the proportion of cells that were treated with TransIT‐TKO, but no siRNA (positive control). Dashed or bold line histograms represent the proportion of cells that were treated with control siRNA or EGFR siRNA, respectively. FITC, fluorescein‐isothiocyanate.

EGFR siRNA leads to a prolonged suppression in the level of EGFR expression and inhibits cell growth in HSC‐2 and SAS cells.  We next examined the duration of siRNA‐mediated decreases in the expression of EGFR. The cells were transfected with siRNA directed against EGFR or its scrambled siRNA. Mock transfected cells were used as controls. The cells were harvested and the expression of EGFR was analyzed by flow cytometry for 8 days post‐transfection. This analysis revealed that the expression levels of EGFR decreased 2 days after siRNA transfection and thereafter remained reduced in both HSC‐2 and SAS cells at least for 8 days (Fig. 2a). Neither the scrambled siRNA nor the mock transfected cells were affected at the level of EGFR expression (data not shown).

Figure 2.

Epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) suppressed the expression of EGFR and inhibited cell growth in HSC‐2 and SAS cells. (a) HSC‐2 and SAS cells were transfected with EGFR siRNA (50 nM), and scrambled siRNA‐transfected or mock transfected cells were used as controls. Mock transfected cells were treated with TransIT‐TKO, but no siRNA. The cells were harvested at the indicated time points and flow cytometric analysis was carried out. Upper panels show the inhibitory effects of EGFR siRNA on EGFR expression levels compared with mock transfected cells by flow cytometric analysis for 8 days post transfection. The lower panels show histograms representing the proportion of cells that were treated with EGFR siRNA (bold line) and mock transfected cells (filled) on day 3 post transfection. (b) HSC‐2 and SAS cells were transfected with 50 nM EGFR siRNA, control siRNA or mock transfected. At the indicated time points, the cells were trypsinized and counted by the trypan blue dye exclusion method in triplicate samples. The results represent the mean ± SEM. The experiment was repeated two additional times with similar results. ▪, mock transfected; ▿, control siRNA (50 nM); •, EGFR siRNA (50 nM). *P < 0.05 in comparison to the control siRNA. PE, phycoerythrin.

To determine the effects of EGFR siRNA on the growth of the two HNSCC cell lines, a series of cell growth experiments was carried out. The growth curves of HSC‐2 and SAS cells showed that treatment with EGFR siRNA, but not with control siRNA, inhibited cell growth over a period of 7 days. The growth inhibitory effects were more significant in SAS cells than in HSC‐2 cells (Fig. 2b).

Enhanced chemosensitivity to cisplatin, 5‐FU or docetaxel by EGFR siRNA in HSC‐2 and SAS cells.  We addressed whether EGFR siRNA would enhance chemosensitivity to cisplatin, 5‐FU or docetaxel in HSC‐2 and SAS cells. Twenty‐four hours after transfection with EGFR siRNA, control siRNA or mock transfection, cells were harvested, transferred to 96‐well plates and treated with cisplatin, 5‐FU or docetaxel for 48 h. The IC₅₀ was determined by MTS assay. Treatment with control siRNA had no effect on the IC₅₀ of either cell line. The cells exposed to EGFR siRNA in the presence of cisplatin, 5‐FU or docetaxel showed a significant decrease in IC₅₀ compared with cells treated with either control siRNA or mock transfection in both cell lines (Fig. 3a). These results indicate that the EGFR siRNA‐mediated decreases in EGFR expression resulted in enhanced tumor cell killing by these chemotherapeutic agents. To evaluate the growth inhibitory effects of EGFR siRNA in combination with cisplatin, 5‐FU or docetaxel in HNSCC cells, we carried out cell growth assays. As shown in Fig. 3b, EGFR siRNA treatment significantly enhanced the growth inhibitory effect of cisplatin, 5‐FU and docetaxel in both cell lines. The control siRNA had either no effect or only a minimal effect.

Figure 3.

(a) Effects of epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) on cisplatin, 5‐fluorouracil (5‐FU) and docetaxel sensitivity in HSC‐2 and SAS cells. In vitro cytotoxicity assays were carried out by MTS assays. Twenty‐four hours after transfection with EGFR siRNA, control siRNA (50 nM) or mock transfection cells were harvested and transferred to 96‐well plates. After treatment with cisplatin, 5‐FU or docetaxel for 48 h, MTS assay was carried out to determine the IC₅₀. The IC₅₀ is the concentration of cisplatin (left), 5‐FU (middle) or docetaxel (right) that caused a 50% reduction of absorbance relative to untreated cells. (b) Growth‐inhibitory effects of treatment with EGFR siRNA in combination with cisplatin, 5‐FU or docetaxel in HSC‐2 and SAS cells. HSC‐2 and SAS cells were treated with 50 nM control or EGFR siRNA for 24 h and were then exposed to cisplatin, 5‐FU or docetaxel (IC₅₀) for an additional 48 h. The results represent the mean ± SEM of three different experiments, and each was carried out in triplicate. *P < 0.05 in comparison to control siRNA. CDDP, cisplatin; DOC, docetaxel.

EGFR siRNA increased cisplatin, 5‐FU and docetaxel‐mediated apoptosis in HNSCC cells in vitro.  To investigate whether the cooperative antiproliferative effect was accompanied by an increase in apoptotic cell death, we examined the effect of EGFR siRNA either with or without chemotherapeutic agents on apoptosis induction in HNSCC cell lines. A flow cytometric analysis was carried out to determine whether the levels of apoptotic fraction were altered in the presence of EGFR siRNA and chemotherapeutic agents in both HSC‐2 and SAS cells. As shown in Fig. 4, the treatment of these cells with 50 nM EGFR siRNA in combination with cisplatin, 5‐FU or docetaxel resulted in a significant increase in sub‐G₁ DNA content in comparison to treatment with chemotherapeutic agents alone. However, the combination therapy had no synergistic apoptosis‐inducing effect on the control siRNA transfected and mock transfected cells.

Figure 4.

Effects of epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) in combination with cisplatin, 5‐fluorouracil (5‐FU) or docetaxel on apoptosis in HSC‐2 (a) and SAS (b) cells. HSC‐2 and SAS cells were transfected with 50 nM control siRNA or EGFR siRNA for 24 h and were then exposed to cisplatin, 5‐FU or docetaxel (IC₅₀) for an additional 48 h. The IC₅₀ is the concentration that caused a 50% reduction of absorbance relative to untreated cells. Apoptotic cells were present in the area indicated by a bar on the left side of each histogram. The number in each panel represents the percentage of apoptotic cells as calculated by a flow cytometric analysis. The data represent one of three different experiments showing similar results. CDDP, cisplatin; DOC, docetaxel.

EGFR siRNA in combination with cisplatin enhanced the antitumor effect of cisplatin in HNSCC xenografts.  To investigate whether the cooperative antitumor effect of EGFR siRNA in combination with chemotherapeutic agents observed in vitro could be reproduced in vivo, we examined the combination effect of EGFR siRNA and cisplatin treatment in BALB/c nu/nu mice bearing SAS tumor xenografts. The tumor volumes in various treatment groups on day 21 were: EGFR siRNA plus cisplatin, 684 ± 86 mm³; control siRNA plus cisplatin, 1158 ± 69 mm³; cisplatin, 1348 ± 48 mm³; EGFR siRNA, 1208 ± 106 mm³; control siRNA, 1656 ± 51 mm³; atelocollagen, 1602 ± 133 mm³; non‐treatment, 1697 ± 107 mm³ (Fig. 5). No significant difference was found in the mean tumor volumes between the non‐treatment group and the control siRNA group. Furthermore, atelocollagen did not affect tumor growth in comparison to the non‐treatment group. EGFR siRNA treatment alone caused tumor growth inhibition comparable to cisplatin treatment. The combination treatment of EGFR siRNA with cisplatin caused a significant inhibition in the tumor growth in comparison to the combination of control siRNA with cisplatin (P = 0.0027) or cisplatin alone‐treated mice (P = 0.0002). Control siRNA had only a minimal effect on tumor growth and on cisplatin efficacy, thus demonstrating the specificity of EGFR siRNA. The animals in all treatment groups appeared healthy and no weight loss or other signs of systemic toxicity were observed in any animals during treatment (data not shown).

Figure 5.

An in vivo study of epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) in combination with cisplatin treatment in SAS human tumor xenografts. The mice were injected subcutaneously into the right flank with 3 × 10⁶ SAS cancer cells. When the tumors reached a volume of approximately 100 mm³, the mice were treated as follows: □, no treatment; ▵, atelocollagen; ○ and •, control and EGFR siRNA (5 µM/50 µL/dose intratumoral injection); ▿, cisplatin (1 µg/g/dose i.p); ▪ and ▴, combination of control or EGFR siRNA (5 µM/50 µL/dose intratumoral injection) and cisplatin (1 µg/g/dose i.p.), at the indicated time point. EGFR and control siRNA were mixed with atelocollagen. The tumor size was measured at the indicated time points and then the tumor volumes were calculated. Each group consisted of six mice. The values represented as mean ± SEM. EGFR siRNA‐treated mice versus cisplatin‐treated mice (no significant difference); EGFR siRNA plus cisplatin‐treated mice versus the cisplatin‐treated mice (*P = 0.0002) or control siRNA plus cisplatin (*P = 0.0027); control siRNA‐treated mice versus no treatment mice (no significant difference); control siRNA plus cisplatin‐treated mice versus cisplatin‐treated mice (no significant difference).

Immunohistochemical analysis of SAS tumor xenografts.  An immunohistochemical analysis of SAS tumors from the different groups of mice was carried out at the end of treatment to evaluate the expression of various biological parameters. The results showed marked differences in the tumors from mice treated with EGFR siRNA plus cisplatin compared with tumors from the mice treated with control siRNA plus cisplatin or cisplatin alone. Inhibition of EGFR expression was observed in the tumors treated with EGFR siRNA and EGFR siRNA plus cisplatin, but not control siRNA and cisplatin. Reductions in tumor cell proliferation (Ki‐67 staining) and in tumor vascularization (CD34 staining) and an increase in tumor cell apoptosis (TUNEL staining) were observed in the tumors treated with EGFR siRNA plus cisplatin (Fig. 6a).

Figure 6.

Effect of epidermal growth factor receptor (EGFR) small interfering RNA (siRNA) in combination with cisplatin on proliferation, angiogenesis and apoptosis in vivo. (a) Immunohistochemical analysis of SAS tumor xenografts (day 21) stained with hematoxylin and eosin (H&E), anti‐EGFR, anti‐Ki‐67 nuclear antigen, anti‐CD34 vessel staining and apoptosis by terminal deoxynucleotidyl transferase‐mediated deoxyuridine triphosphate brotin nick‐end labeling (TUNEL) (magnification, ×400). (b) The quantification of proliferation index (Ki‐67), the microvessel density and the apoptosis index (TUNEL). All quantitative data are presented as mean ± SEM. *P < 0.05 in comparison to cisplatin alone.

Quantitative analysis of tumor cell proliferation, apoptosis and vessel staining showed a significant difference between treatment with EGFR siRNA plus cisplatin and treatment with cisplatin alone (P < 0.05). Treatment with EGFR siRNA plus cisplatin resulted in a 1.5‐fold decrease in tumor cell proliferation, a 1.7‐fold increase in tumor cell apoptosis, and a 1.5‐fold decrease in vessel staining, in comparison to cisplatin treatment (Fig. 6b). Cell proliferation, apoptosis and angiogenesis in tumors from control siRNA‐ or atelocollagen‐treated mice did not differ significantly from those observed in the tumors from non‐treatment mice (data not shown). These results indicated that the antitumor activity of combined treatment with EGFR siRNA and cisplatin was a result of decreased tumor cell proliferation, decreased tumor vascularization, and increased tumor cell apoptosis in comparison to those treated with cisplatin alone.

Discussion

Patients with HNSCC may receive single‐agent or combination chemotherapy, including cisplatin, 5‐FU and docetaxel. Despite the fact that many different chemotherapeutic agents and regimens have been developed to manage HNSCC, overall there has been no significant increase in the survival of patients. Resistance to conventional chemotherapy and dose‐limiting toxicities remain major issues in the efforts to improve survival rates. Therefore, development of novel chemotherapeutic strategies is necessary if the outcome for patients with HNSCC is to improve.

Epidermal growth factor receptor plays an important role in malignant transformation and tumor growth through the inhibition of apoptosis, cellular proliferation, angiogenesis and metastasis.^{( 10} , ¹¹ , ¹² , ^{13 )} Overexpression of EGFR has been reported in a variety of malignant disorders, including head and neck cancer.^{( 5} , ⁶ , ⁷ , ⁸ , ^{9 )} Therefore, several methods for inhibiting EGFR expression or activation, such as antisense oligonucleotides, monoclonal antibodies and EGFR‐specific tyrosine kinase inhibitors, are currently under investigation in preclinical and clinical settings against various types of cancer.^{( 4} , ¹⁴ , ^{25 )} Although promising results have been observed in a preclinical setting, their efficiency is far from satisfactory when these agents have been administrated as monotherapy to patients.^{( 15} , ^{16 )} Therefore, combination therapies with EGFR inhibitors and chemotherapeutic regents are currently receiving considerable interest in the treatment of cancers.

In the present study, we used the RNA interference technique, using siRNA to downregulate EGFR expression. Our results demonstrated that EGFR siRNA, a duplex of 21 nt, effectively downregulated up to 90% of EGFR expression in HNSCC cells at day 3 after transfection (Fig. 2). Synthetic double‐stranded RNA longer than 30 nt will induce an interferon response in mammalian cells, but siRNA, a duplex of 21–23 nt, avoids the induction of such a response, and it is also sequence‐specific to target genes.^{( 17} , ¹⁸ , ^{26 )} When RNA interference technology is used for cancer therapy, one theoretical benefit of the target‐specific therapy is a lower risk of systemic side‐effects. However, in addition to the specific gene‐silencing effects of siRNA, others have shown that activation of an interferon response caused by upregulation of global interferon‐stimulated genes could be induced by 21 nt siRNA in mammalian cells.^{( 27} , ^{28 )} To limit the risk of inducing an interferon response, siRNA should be used at the lowest effective dose.

Although siRNA inhibits the expression of mRNA as antisense oligonucleotides, the intracellular cascade is quite different. From studies directly comparing siRNA with antisense oligonucleotides, siRNA is considered to be more efficient than antisense‐based approaches in its resistance to nucleases and in inhibiting target gene expression in vitro and in vivo.^{( 22} , ^{23 )} However, this point should be more precisely investigated, as a report has shown that siRNA directed against EGFR was less effective than the better‐characterized antisense oligonucleotide when administered in combination with several chemotherapeutic agents in a PC‐3 prostate cell line.^{( 29 )} In addition, siRNA may have advantages over monoclonal antibodies for silencing oncogenes such as EGFR. Although the effect of monoclonal antibody is limited to the time it remains present in the vicinity of the tumor cells at a therapeutic level, siRNA leads to prolonged inhibition of target protein expression in the parent as well as its progeny cells.^{( 30} , ^{31 )} Taken together, while issues related to induction of an interferon response by siRNA remain, these cumulative studies suggest that RNA interference technology targeting cancer‐associated gene products (including EGFR) may be powerful therapeutic agents for cancer.

To date, several preclinical studies have reported EGFR inhibitors to have an additive or synergistic antitumor effect when administered in combination with chemotherapeutic agents in a variety of cancer models, including HNSCC.^{( 32} , ³³ , ³⁴ , ^{35 )} Consistent with these reports, we demonstrated that the downregulation of EGFR expression by siRNA significantly enhanced the growth‐inhibitory effects and chemosensitivity of cisplatin, 5‐FU and docetaxel in HNSCC cells (Fig. 3). Furthermore, we also demonstrated that EGFR siRNA significantly increases the number of apoptic cells in response to treatment with these chemotherapeutic agents (Fig. 4). Our results thus suggest that downregulation of EGFR may play an important role in enhancing apoptosis induced by cisplatin, 5‐FU and docetaxel. However, the mechanisms responsible for this enhanced chemosensitivity by downregulation of EGFR have not yet been fully clarified. Several reports have shown that EGFR activation regulates apoptosis in different experimental systems, and the inhibition of EGFR was associated with a reduction in tumor cell growth and an induction of apoptosis.^{( 35} , ^{36 )} In contrast, other reports have shown that siRNA directed against wild‐type EGFR had no effect on cell growth or apoptosis.^{( 37 )} Chemotherapeutic agents are likely to induce a complex cascade of molecular and biochemical changes in the expression of myriad genes associated with cellular protection or cell death response.^{( 38} , ^{39 )} For example, the anti‐apoptotic Bcl‐2 family of proteins is known to regulate apoptotic events in response to chemotherapeutic agents,^{( 40 )} and EGFR‐mediated signaling pathways are also associated with regulating the expression and activation of Bcl‐2 proteins.^{( 41 )} These apoptotic pathways initiated by chemotherapeutic agents and stimulated by inhibition of EGFR may overlap with each other and the combined inhibition of these two pathways may result in enhanced chemosensitivity. The interactions between chemotherapeutic agents and various molecules that regulate tumor cell growth and apoptosis in the EGFR‐mediated signaling pathways are under study. As with chemotherapy, various studies have also shown an enhancement of tumor response to radiation therapy combined with EGFR inhibitors.^{( 42} , ^{43 )} Taken together, these studies suggest the possibility that EGFR ligands might become survival factors rather than growth factors, tipping the balance further towards apoptosis, in the presence of cellular damage induced by chemotherapy or radiation therapy, and repair of the cellular damage might be blocked by inhibition of the EGFR‐mediated signaling pathway.^{( 42} , ⁴³ , ^{44 )}

Based on the findings of the present in vitro experiments, we established tumors by subcutaneous injection of HNSCC cells into nude mice, and also examined whether EGFR siRNA therapy enhances the cytotoxic effect of cisplatin on the growth of HNSCC tumors. Cisplatin was used in the present study because it was more efficient than 5‐FU and docetaxel in inhibiting cell growth in in vitro studies. Consistent with our in vitro studies (Fig. 3b), we demonstrated that treatment with EGFR siRNA in combination with cisplatin significantly inhibited the growth of HNSCC tumors in comparison to treatment with cisplatin alone in a HNSCC xenograft model. The antitumor effect of EGFR siRNA alone was comparable to cisplatin treatment (Fig. 5).

A histological examination of HNSCC xenograft tumors from EGFR siRNA plus cisplatin‐treated mice showed dramatic changes in several markers (Ki‐67, TUNEL and CD34) that could be responsible for the enhanced antitumor effects (Fig. 6). Our results are consistent with those of several studies that have demonstrated that EGFR inhibitors combined with commonly used chemotherapeutic agents induced tumor cell apoptosis in various human xenograft tumor models.^{( 34} , ⁴⁵ , ⁴⁶ , ^{47 )} The decreased CD34 vessel staining observed in HNSCC xenografts suggests that anti‐angiogenic mechanisms may thus be associated with a downregulation of EGFR by siRNA treatment. To date, several reports have shown that the EGFR signaling pathway plays an important role in the regulation of angiogenesis in local tumor growth, and that blockade of the EGFR signaling pathway has been shown to inhibit tumor‐induced angiogenesis in various human tumor models.^{( 48} , ⁴⁹ , ^{50 )} These results suggest that the decrease in tumor cell proliferation and angiogenesis, and the increase in apoptosis, accompanied by downregulation of EGFR expression, were likely responsible for the enhanced antitumor effects observed in EGFR siRNA plus cisplatin‐treated tumors. Further studies are necessary to clarify the mechanisms of action for these chemotherapeutic agents against the EGFR signaling pathway in regulating tumor cell proliferation, angiogenesis and apoptosis.

The development of an in vivo delivery system for siRNA is a critical factor in its use as a therapeutic agent for cancer. Atelocollagen is a natural biomaterial produced from bovine type I collagen and it is neither antigenic nor toxic in animals because antigenic telopeptides attached to both ends of the collagen are eliminated by pepsin digestion.^{( 24} , ^{51 )} In addition, atelocollagen is soluble at 4°C but solidifies to refibrillation at 37°C and is gradually biodegraded.^{( 52 )} To date, several reports have described that the potency of nucleic acid‐based gene materials, such as plasmid DNA and antisense oligonucleotides, was greatly enhanced through the use of atelocollagen in vitro and in vivo.^{( 24} , ⁵³ , ⁵⁴ , ^{55 )} Recently, Takei et al. reported that fluorescein‐labeled and radiolabeled siRNA mixed with atelocollagen stabilized the tumors for at least 1 week, and intratumoral injection of atelocollagen in complex with siRNA against vascular endothelial growth factor showed dramatically suppressed tumor angiogenesis and tumor growth in a PC‐3 sc xenograft model.^{( 56 )} More recently, Takeshita et al. reported that the systemic administration of atelocollagen in complex with siRNA targeting human enhancer of zeste homolog 2 into a mouse model of bone metastasis showed efficient inhibition of tumor growth without inducing any side‐effects.^{( 57 )} In the present study, we demonstrated that intratumoral injection of atelocollagen in complex with EGFR siRNA downregulated the expression of EGFR and enhanced the antitumor effect of cisplatin in a HNSCC xenograft model. These results suggest that an atelocollagen‐mediated siRNA delivery system is a reliable method for gene therapy using siRNA.

In conclusion, we showed that siRNA targeting EGFR efficiently downregulated EGFR expression with specificity and inhibited cell growth in HNSCC. Treatment with EGFR siRNA in combination with cisplatin, 5‐FU and docetaxel showed enhanced chemosensitivity with a significant increase in apoptosis. Furthermore, we also showed that EGFR siRNA delivered by atelocollagen enhanced the antitumor effect of cisplatin in an HNSCC xenograft model. These cumulative results strongly suggest that EGFR siRNA combined with cisplatin, 5‐FU and docetaxel may thus be a feasible and complementary strategy to enhance the effects of chemotherapy in patients with HNSCC.