Synergistic inhibition of head and neck tumor growth by green tea (−)-epigallocatechin-3-gallate and EGFR tyrosine kinase inhibitor

Cell lines

Five cell lines (Tu177, Tu212, 886LN, SQCCY1 and 38) were selected for our study to represent different SCCHN sites. Tu177 and Tu212 cell lines established from a laryngeal and hypopharyngeal tumor, respectively, were provided by Dr. Gary L. Clayman (University of Texas M.D. Anderson Cancer Center, Houston, TX). The 886LN cell line was derived from a lymph node metastasis of squamous cell carcinoma of the larynx. SQCCY1 and 38 cell lines derived from oral cavity and tonsil fossa, respectively, were obtained from Dr. Shi-Yong Sun (Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA). All cells were maintained in DMEM/F12 (1:1) medium supplemented with 5% heat-inactivated fetal bovine serum (FBS) and antibiotics (streptomycin, penicillin G and amphotericin B) in a 37°C, 5% CO₂ humidified incubator. For 3 to 5 days of treatment, medium was changed on day 3.

Reagents

EGCG (Sigma Chemical, St. Louis, MO) and erlotinib (Genentech, South San Francisco, CA) were dissolved in autoclaved water and dimethyl sulphoxide (DMSO), respectively, as stock solutions for in vitro studies. Since there is substantial precipitate in DMSO-dissolved erlotinib, we harvested the supernatant after spinning down the precipitate for use as a stock solution. The reagents were further diluted in DMEM/F12 medium immediately before use. The final concentration of DMSO was less than 0.1%. For in vivo studies, EGCG was dissolved in autoclaved water, while erlotinib was dissolved in 1% Tween-80 (Sigma Chemical). Superoxide dismutase (SOD) was purchased from Sigma Chemical.

Cell growth inhibition assay

To test the effects of single agent EGCG and erlotinib on cell growth of SCCHN, sulforhodamine B (SRB) cytotoxicity assays were adapted from Skehan et al.²⁵ Cells maintained in medium with 5% FBS were seeded in 96-well plates at a density of 4,000 cells/well overnight prior to drug treatment. Afterwards, drugs were added as single agents in various concentrations (0–100 μM for EGCG, 0–40 μM for erlotinib), followed by incubation at 37°C and 5% CO₂ for 72 hr. Cells were fixed for 1 hr with 10% cold trichloroacetic acid. Plates were washed 5 times in water, air-dried and then stained with 0.4% SRB for 10 min. After washing 4 times in 1% acetic acid and air-drying, bound SRB was dissolved in 10 mM unbuffered Tris base (pH 10.5). Plates were read in a microplate reader by measuring absorbance at 492 nm. The percent survival was then calculated based upon the absorbance values relative to untreated samples. The experiment was repeated 3 times.

Combination index assay

The interaction between EGCG and erlotinib was analyzed using the CalcuSyn software program (Biosoft, Ferguson, MO) which is based on the Chou and Talalay method.²⁶ In brief, cells were seeded in 96-well plates at a density of 4,000 cells/well overnight prior to drug treatment. The next day, 1:2 serial dilutions of the drugs were prepared for EGCG, erlotinib and their combination based on their IC₅₀, and added to the cells. After incubation for 72 hr at 37°C and 5% CO₂, the plates were subjected to SRB assay. Data from the SRB assay were expressed as the fraction of cells with growth affected (FA) in drug-treated vs. untreated cells. The CI was calculated using CalcuSyn software. A CI value of >1 is defined as antagonism, equal to 1 as additivity and <1 as synergy. The experiment was repeated 3 times.

Flow cytometry analysis

Cell cycle arrest and apoptosis were analyzed in 2 cell lines, Tu212 and Tu177. Cells were treated with EGCG (30 μM), erlotinib (0.5 μM) or their combination (EGCG 30 μM + erlotinib 0.5 μM) for certain time points, then trypsinized and washed in cold 1× PBS. For cell cycle analysis, cells were fixed in 70% ethanol at 4°C for 2 hr, followed by a wash with 1× PBS. Cells were then stained with PI/RNase staining buffer (BD Pharmingen, San Diego, CA) for 15 min at room temperature. For Annexin V and 7-amino-actinomycin (7-AAD) binding assay, after trypsinizing and washing with cold PBS, the cells were resuspended in 1× Annexin binding buffer (BD PharMingen), and then stained with Annexin V-phycoerythrin (Annexin V-PE; BD PharMingen) and 7-AAD (BD PharMingen) for 15 min at room temperature. The stained samples for both cell cycle and apoptosis assays were measured using a fluorescence-activated cell sorting (FACS) caliber bench-top flow cytometer (Becton Dickinson, Franklin Lakes, NJ). The data for cell cycle arrest were analyzed using Cell Quest software (Becton Dickinson, Franklin Lakes, NJ). FlowJo software (Tree Star, Ashland, OR) was used for apoptosis analysis. The experiments were repeated 3 times independently.

Immunofluorescence

Tu212 cells were serum-starved overnight, then pre-exposed to 10 μM EGCG, 0.5 μM of erlotinib or the combination for 4 hr, followed by treatment with 100 ng/ml of EGF for 15 min, or cells were pre-exposed to EGCG for 7 hr and in the last 3 hr cells were treated with EGF for 3 hr.

Cells grown on coverslips were fixed with warm PHEMO buffer [68 mM PIPES, 25 mM HEPES (pH 6.9), 15 mM EGTA, 3 mM MgCl₂, 10% (vol/vol) DMSO] containing 3.7% formaldehyde, 0.05% glutaraldehyde and 0.5% Triton X-100 for 10 min and incubated with primary antibody and fluor-conjugated secondary antibodies. Images were taken with Zeiss LSM510 META confocal microscope.

Nude mouse xenograft model

The animal experiments were approved by the Animal Care and Use Committee of Emory University. Twenty-five nude mice (athymic nu/nu, Taconic, NY), aged 4 to 6 weeks (about 20 g weight), were randomly divided into 4 groups. Each group was orally gavaged for 7 days with vehicle control (1% Tween 80, n = 6), EGCG (125 mg/kg, n = 6), erlotinib (50 mg/kg, n = 6) or the combination (n = 7) of EGCG (125 mg/kg) and erlotinib (50 mg/kg) using a blunt tipped 20G 1.1/2 needle (Popper and Sons, New Hyde Park, NY) before inoculation of 2 × 10⁶ Tu212 cells by s.c. injection into the right flank. The animals were continuously administered the agents 5 days a week. The tumor size was measured 3 times a week. The tumor volume was calculated using the formula: V = π/6 × larger diameter × (smaller diameter)², as reported previously.²⁷ Growth curves were plotted using average tumor size within each experimental group at the set time points. The median time from the tumor cell injection until the average tumor size (V) reached 500 mm³ in each group was evaluated by the Kaplan–Meier curve. The whole group of mice was sacrificed once the size of any tumor in that group reached 2 cm in diameter.

Western blot analysis

Whole cell lysates were extracted from drug-treated cells using lysis buffer. Fifty microgram protein was separated on 8–15% SDS-PAGE, transferred onto a PVDF membrane (Millipore, Bedford, MA) and immunostained with specific antibodies including rabbit anti-pEGFR and EGFR (Santa Cruz, CA), rabbit anti-pAKT and AKT (Cell Singling, Beverly, MA), mouse anti-caspase 3 (Imgenex, San Diego, CA), mouse anti-caspase 8, anti-caspase 9 and anti-PARP (Cell Singling, Beverly, MA). Rabbit anti-G3PDH or mouse anti-β-actin antibody (Trevigen, Gaithersburg, MD) was used as a sample loading control. Immunostained protein bands were detected with an enhanced chemiluminescence kit (Amersham, Buckinghamshire, UK). The experiments were repeated 3 times.

Protein (100 μg) from each sample of nude mouse xenograft tissue was subjected to immunoblotting analysis using corresponding antibodies. The intensity of Western blotting was analyzed by densitometric image analysis system (UVP BioImaging System, Upland, CA).

For immunoprecipitation studies, about 300 μg of lysate was precleared with recombinant protein G-agarose (GIBCOBRL, Carlsbad, CA) for 4 hr at 4°C, then incubated with 4 μg of antibody against the N-terminus of EGFR (528, SC-120, Santa Cruz, CA) or nonimmunized mouse IgG precomplexed with protein G-agarose overnight. The membrane was probed with anti-ubiquitin antibody (Cell Signaling, Beverly, MA).

Immunohistochemistry and TUNEL assay

Immunohistochemistry analysis for Ki-67 staining on formalin-fixed, paraffin-embedded nude mouse xenograft tissue was performed using the R.T.U. Vectastain kit following the standard manufacture's protocol (Vector Laboratories, Burlingame, CA). Tissue sections were incubated with mouse anti-human Ki-67 (prediluted; Biomeda, Foster, CA) overnight at 4°C. The slides were stained with 3,3′-diaminobenzidine (DAB) (Sigma Chemical) and counterstained with hematoxylin (Vector Laboratories). TUNEL assay was performed by immunofluorescence using the same specimens as above following the procedure provided by the manufacturer (Promega, Madison, WI). The slides were counterstained with DAPI (Vector Laboratories).

Both Ki-67 staining and fluorescent signals from the TUNEL assay were visualized under Zeiss LSM 510 confocal microscopy. Their images were analyzed by MetaMorph software (Universal Imaging, Downingtown, PA) to measure the intensity or count the positive cells. Ten areas were randomly selected from each slide for measurement. For Ki-67 staining, we measured the intensity of positive cells by counting the absolute number of pixels. For TUNEL assay, we counted the total cell number and the positive cell number in the same area; the result was presented as the positive cell number per 1,000 total cells.

Statistical analysis

The two-tailed and paired t tests were used for the cell cycle and apoptosis assays. The two-tailed unpaired t test was used for analyzing the expression level of pEGFR and pAKT in xenograft tissue. The Wilcoxon rank-sum test was applied to evaluate the significance of differences in tissue necrosis, Ki-67 expression and TUNEL assay. The method of repeated measures ANOVA with between-subject factors was used to evaluate the significance of tumor cell growth inhibition among each group. Kaplan–Meier curves and log-rank test were employed for tumor progression analysis.

Growth of SCCHN cells

To study the sensitivity of SCCHN to EGCG and erlotinib, we initially examined 5 SCCHN cell lines (Tu212, Tu177, 886LN, SQCCY1 and 38), of which Tu212, Tu177, 886LN express comparable amounts of total EGFR protein and activated EGFR at basal level. Tu212 and Tu177 are sensitive to EGFR-TKI inhibition³; SQCCY1 is sensitive to apoptosis induced by inhibition of the AKT pathway²⁸; and 38 is a chemoresistant cell line.²⁹ Cells were treated with EGCG (0–100 μM) or erlotinib (0–40 μM) for 72 hr. A cell growth inhibition assay (SRB) showed that both single agent EGCG and erlotinib inhibited growth of the 5 SCCHN cell lines in a dose-dependent manner (Figs. 1a and 1b). The effect of combined treatment with EGCG and erlotinib on growth inhibition of SCCHN cells was determined by combination index (CI) assay. The concentrations were based on IC₅₀ derived from the growth inhibition assay for single agents with a series of 2-fold reduction or escalation. The same concentration used for the single drug was also applied for the combination. The CI indicated that the two-drug combination of EGCG and erlotinib synergistically inhibited cell growth in 4 of 5 tested SCCHN cell lines (Fig. 1c). A synergistic effect on cell growth inhibition was also observed at relatively high concentrations in 886LN cells. Therefore, single doses of erlotinib and EGCG at their IC₅₀s were applied for the remaining in vitro studies.

Cell cycle progression and induction of apoptosis

The effects of the two-drug combination on cell cycle progression were investigated in 2 representative SCCHN cell lines, Tu212 and Tu177. Cells were treated for 24 and 48 hr with 30 μM EGCG and 0.5 μM erlotinib (Fig. 2a). Compared to the control group, erlotinib alone induced significant cell cycle arrest at G0/G1 phase in both Tu212 and Tu177 cells as early as 24 hr after treatment. EGCG alone exerted no effect on cell cycle arrest. The combined treatment induced additive cell cycle arrest compared to that induced by EGCG or erlotinib alone in either cell line. The two-drug combination did not affect cell cycle changes at either S or G2/M phase. The apoptotic effect of the combination treatment was examined by treating Tu212 and Tu177 cells for 72 and 96 hr. Single agent EGCG or erlotinib induced moderate apoptosis, but the two-drug combination resulted in time-dependent significantly greater apoptotic cell death (Fig. 2b, p < 0.05).

Targeting phosphorylation of EGFR and AKT pathways

To elucidate the molecular mechanisms of EGCG and erlotinib treatment in SCCHN, we examined whether the drugs could modulate pEGFR and pAKT levels. Tu212 cells were treated with EGCG (30 μM), erlotinib (0.5 μM) or their combination for 30 min, 6, 24 and 72 hr. Erlotinib alone inhibited the phosphorylation of both pEGFR and pAKT after 30 min and retained this effect until 72 hr. EGCG alone clearly suppressed pAKT expression but did not substantially alter pEGFR levels until 72 hr. However, the combination of EGCG and erlotinib further downregulated pAKT after 30 min and pEGFR after 72 hr (Fig. 3a).

A study by Hou et al. suggested that SOD can stabilize EGCG in vitro by blocking auto-oxidation, resulting in greater inhibition of cell growth by EGCG.³⁰ Therefore, we examined the effect of SOD on the efficacy of EGCG and its synergy with erlotinib in inducing apoptosis. Tu212 cells were treated with EGCG, erlotinib or their combination for 3 and 5 days in the presence or absence of SOD (5 U/ml). Figure 3b shows that (i) inhibition of pEGFR, pERK and pAKT by erlotinib was associated with marked increases of cleaved caspase 9, PARP and caspase 3; (ii) in the presence of SOD, EGCG induced greater reduction of pEGFR on day 3 and total EGFR on day 5 than that without SOD; (iii) the combination treatment induced substantially greater reduction of pEGFR, pERK, total EGFR and AKT levels associated with greater increases in activation of caspase 3 than the treatment with the single agents. This effect was greater in the presence vs. the absence of SOD and became more pronounced by day 5 relative to day 3. We did not observe noticeable changes in the activation of caspase 8. Similar results were obtained with other SCCHN cell lines (data not shown).

Depletion of EGFR protein

To further explore the mechanism by which combination treatment affects EGFR and other cellular signaling pathways, we tested a hypothesis that erlotinib inhibits phosphorylation of EGFR, immobilizing it at the plasma membrane, as seen in SCCHN,³¹ while EGCG enhances EGFR internalization and degradation; thus the combination of EGCG and erlotinib leads to the ultimate depletion of total EGFR protein. We performed parallel immunofluorescence (Fig. 4a) and Western blot analyses (Fig. 4b) using Tu212 cells in the presence or absence of EGF to examine whether EGCG affects EGF-induced EGFR activation, internalization and degradation.^32,33

Immunofluorescence staining of EGFR (Fig. 4a) shows that at the basal level, EGCG alone induced EGFR internalization, shown as diffusive cytoplasmic staining which became more pronounced by 7 hr of treatment. While erlotinib alone stabilized EGFR at the membrane, it did not prevent EGFR internalization induced by EGCG when in combination treatment for 7 hr. EGFR was internalized when cells were briefly exposed to EGF; clusters of weak staining in the cytoplasm were detected after 3 hr. In the presence of EGF, internalization of EGFR was detected in cells treated with EGCG, whereas erlotinib consistently prevented EGFR internalization even after 3 hr of exposure, as shown by the membranous staining of EGFR. However, when erlotinib was combined with EGCG, most EGFR remained in the cytoplasm with a markedly reduced intensity by 3 hr of treatment compared to the level at 15 min, suggesting that EGFR fails to recycle back to plasma membrane and may undergo degradation.

Western blot analysis (Fig. 4b) revealed that at basal level, EGCG alone or in combination with erlotinib reduced pEGFR, pERK, pAKT and total EGFR protein levels. Similarly, erlotinib inhibited pEGFR, pERK and pAKT but not total EGFR levels. EGF induced intense activation of pEGFR (tyr1173, tyr 1045) and pERK. EGF also induced significant downregulation of EGFR to undetectable levels after 3 hr of treatment. EGCG had no effect on activation of EGFR (tyr 1173 and tyr1045), pERK and pAKT by EGF, but specifically augmented ligand-induced downregulation of total EGFR protein, whereas erlotinib substantially abrogated the activation of pEGFR, pERK and pAKT by EGF without altering the total levels of EGFR protein. Erlotinib in combination with EGCG induced greater inhibition of pEGFR (tyr1173 and tyr1045), pERK and pAKT and greater downregulation of EGFR protein than erlotinib alone.

Ubiquitin-dependent degradation of EGFR

We consistently observed that EGCG induced downregulation of EGFR protein expression, either in short-term serum-free conditions (Fig. 4b) or in long-term culture with serum (Fig. 3b) in Tu212 and other cell lines (data not shown). The presence of SOD in all treatments should prevent loss of EGFR by auto-oxidation. Hence, we asked whether EGCG could induce ubiquitin-dependent degradation of EGFR. Immunoprecipitation revealed that with equal amounts of EGFR precipitated from the lysate, cells treated with EGCG displayed a dose-dependent increase in ubiquitinated EGFR compared to the control (Fig. 4c). When proteasomal activity was inhibited by MG132, a specific proteasome inhibitor, a dose-dependent increase in ubiquitinated EGFR was detected in cells treated with EGCG. Furthermore, ubiquitinated EGFR induced by EGF was augmented in the presence of EGCG, or EGCG with erlotinib compared to the level suppressed by erlotinib alone. Western blot analysis of the lysates for immunoprecipitation revealed that, consistent with the observations shown in Figure 4b, EGCG alone indeed downregulated EGFR and reduced protein levels even further when combined with EGF. In the presence of MG132, total EGFR levels appeared to be increased by EGCG compared to that by MG132 alone.

Growth of mouse xenograft tumors

The antitumor efficacy of combined treatment with EGCG and erlotinib was determined in xenograft mice bearing Tu212 cells, a known tumorigenic cell line. We pretreated mice with vehicle control, EGCG alone, erlotinib alone or their combination for 7 days to imitate an intrinsically preventive environment in the mouse body before injection of Tu212 tumor cells. The animals were continuously given the agents after the tumor cell injection. The tumor growth data indicated that, although EGCG and erlotinib as single agents moderately inhibited tumor growth, no statistical significance was reached compared to the control (p = 0.14, p = 0.07, respectively). However, the combination of EGCG and erlotinib significantly suppressed tumor growth compared to the control (p = 0.006), the EGCG alone treated (p = 0.02) or the erlotinib alone treated group (p = 0.01) (Fig. 5a). Mice treated with EGCG and two-drug combination were observed to be more energetic and active than those in the other 2 groups.

Kaplan–Meier curves revealed that the median times to reach a tumor size of 500 mm³ were 19.5, 25 and 27 days, and not reached at the time of termination for the control, EGCG alone, erlotinib alone and the combined treatment, respectively (Fig. 5b). Log-rank tests showed that this time delay induced by the combined treatment was statistically significant compared to control (p = 0.005), EGCG alone (p = 0.001) or erlotinib alone (p = 0.03). There is a pattern of time delay for the tumor volume to reach 500 mm³ in both the EGCG and erlotinib treatment groups relative to control, but this is not statistically significant, p = 0.5 and p = 0.1, respectively.

Ki-67 expression and apoptosis in xenograft tumors

Expression of Ki-67 and TUNEL assays revealed that the combination treatment resulted in greater inhibition of proliferation and in induction of apoptosis in the xenograft tumors (Fig. 5c). There were statistically significant differences in mean Ki-67 expression between the combination and the control groups (p = 0.03) and between the combination group and the erlotinib group (p = 0.05). Induction of apoptosis by the combination treatment was significantly greater than that by the control group (p = 0.04).

Phosphorylated EGFR and AKT in xenograft tumors

Protein extracted from the xenograft tumor tissues was examined to evaluate whether the inhibition of pEGFR and pAKT by treatment with EGCG and erlotinib in vitro is also seen in vivo. We found that erlotinib alone, but not EGCG alone, significantly suppressed pEGFR levels compared to the control treatment (p = 0.001), and the combined treatment yielded significantly lower mean pEGFR levels than that in the control group (p = .002, Figs. 6a and 6b). Both EGCG and erlotinib moderately inhibited pAKT, the combined treatment further reduced its levels but this was not statistically significant (Figs. 6c and 6d).