Abstract
Background: The epidermal growth factor receptor (EGFR, ErbB-1, HER-1) is overexpressed in many epithelial tumors, particularly head and neck squamous cell carcinomas (HNSCCs) and is related to poor prognosis. For non-small cell lung cancer (NSCLC) it was found that activating mutations in the kinase domain of the receptor predicted a high response-rate to EGFR-specific kinase inhibitors. The goal of the present study was to investigate potential sequence changes of EGFR in HNSCC cells and to determine their possible role in tumor biology. Materials and Methods: The whole EGFR coding sequence of eleven previously well-characterized HNSCC cell lines was determined by RT-PCR sequencing. The response of the cells to the kinase inhibitor AG1478 and the monoclonal anti-EGFR antibody cetuximab was evaluated by cell cycle and Western blot analysis. Results: None of the cell lines exhibited EGFR mutations. However, 4 out of the 11 (36%) cell lines harboured the K497 polymorphism in the receptor. The R497 cell lines were more frequently (71%) derived from N+ tumors than the K497 cell lines (25%), whereas the K497 cells, although not reaching significance, appeared on average to be more sensitive to inhibitor treatment. This effect was particularly pronounced in the AG1478-treated tumor cells and was associated with the level of extracellular-signal regulated kinase-1/2 phosphorylation which appeared more efficiently inhibited in the cell lines exhibiting the K497 EGFR polymorphism. Conclusion: EGFR mutations are a rare event in HNSCC cell lines and, consistent with previous studies the EGFR codon 497 polymorphism could play a significant role in HNSCC disease and therapy response.
The epidermal growth factor receptor (EGFR, ErbB-1, HER-1) is associated with disease progression and reduced prognosis of many epithelial tumors (1). Head and neck squamous cell carcinomas (HNSCCs) overexpress EGFR in up to 100% of cases. For this reason, the EGFR has become a major target in cancer therapy (2). In recent years, monoclonal antibodies such as cetuximab (Erbitux®) and panitumumab (Vectibix®) that target the extracellular domain of the receptor, aiming to block ligand binding, came into broad use in treating EGFR-expressing tumors. Similarly, small molecule kinase inhibitors such as gefitinib (Iressa™) and erlotinib (Tarceva®) were directed against the intracellular kinase domain of the receptor destined to block receptor autophosphorylation by competing with ATP (1). Interestingly, for non-small cell lung cancer (NSCLC) which also overexpress EGFR it was found that the presence of activating kinase domain mutations in the EGFR rendered the cells susceptible to treatment with kinase inhibitors (3). Similarly, polymorphisms in the EGFR gene were found to be associated with the clinical course of the disease as well as therapy response in different epithelial tumors such as colorectal, renal, prostate, breast and esophageal as well as head and neck carcinomas (4-12). Therefore, subpopulations of patients harbouring specific sequence aberrations in the EGFR should benefit from treatment with kinase inhibitors or other anti-EGFR targeted therapies. Hence, it is of clinical relevance to identify those patients and offer them a suitable therapy scheme. The aim of the present study was to investigate potential sequence changes in previously well-characterized EGFR-overexpressing HNSCC cell lines. The sequence results were investigated in relationship to the cellular response to inhibition with cetuximab and the EGFR-specific kinase inhibitor AG1478.
Materials and Methods
Cell lines and cell culture. The squamous cell carcinoma cell lines UM-SCC-1, -3, -4, -14A, -22B and -27, and UT-SCC-24A and -26A were kindly provided by T. E. Carey (University of Michigan, MI, USA) and R. Grénman (University of Turku, Finland) respectively (13). The UMB-SCC-745, -864 and -969 (University of Marburg) cell lines were derived from tumors of the oropharynx, tongue and pharynx. All 11 HNSCC cell lines have been previously well-characterized (Table I) (14-16). The cells were grown in DMEM supplemented with 10% fetal calf serum (FCS) in the presence of penicillin and streptomycin.
Sequence analysis. The total RNA was generated with a RNeasy MIDI kit from Qiagen (Qiagen GmbH, Hilden, Germany) according to the manufacturer's instructions. Reverse transcription was conducted in the presence of an oligo(dT) primer to generate full length cDNA. PCR amplification (96°C, 10 min then 42 cycles of 96°C, 1 min; 55°C, 2 min; 72°C, 5 min followed by 72°C, 15 min) was performed to recover the complete EGFR coding region, using the EGFR-specific primers (f=forward, r=reverse): 5′-AGACCGGACGACAGGCCA-3′ (109, f), 5′-AACGCCACAACCACCGC-3′ (164, f), 5′-CCCCGTAATTATGTGGTGACAGATC-3′ (1132, f), 5′-CGGCAGG ACCAAGCAACA-3′ (1527, f), 5′-CAGCCTGAACATAACATCCTTG-3′ (1569, f), 5′-CATCTGCCTCACCTCCACCG-3′ (2583, f), 5′-CAAGGATGTTATGTTCAGGCTG-3′ (1569, r), 5′-CAGCAGGCGGCACACG-3′ (2565, r), 5′-CGGTGGAGGTGAGGCAGATG-3′ (2583, r), 5′-CCCGTAGCTCCAGACATCACTC-3′ (2928, r), 5′-CTGCTGTGGCTTGGTCCTG-3′ (3927, r), 5′-AATGTGCCCGAGGTGGAAGTA-3′ (4030, r). The number in brackets indicates the position of the primer in the published human wt EGFR sequence (NM_005228 Genbank NCBI, NIH, Bethesda, MD, USA). The PCR products were gel purified (Qiagen) and sent for sequencing (4base lab GmbH, Reutlingen, Germany).
Treatment of cells with AG1478 and cetuximab. AG1478 (Tyrphostin) was purchased from Calbiochem (San Diego, CA, USA). Cetuximab was kindly provided by Merck (Merck Pharma GmbH, Darmstadt, Germany). The HNSCC cells were incubated with 10 μg/ml AG1478 or 100 μg/ml cetuximab for 13 h at 37°C, 5% CO2.
Antibodies, SDS-PAGE and Western blot analysis. Antibodies specific for EGFR (1005) and actin (C-11) as well as all the secondary horseradish peroxidase (HRP)-coupled antibodies were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Those for phospho-EGFR (9H2) was from Calbiochem, Erk-1/2 (extracellular-signal regulated kinase-1/2) from Upstate (Temecula, CA, USA) and phospho-Erk-1/2 from Sigma (St. Louis, MO, USA). SDS-PAGE and Western blot analysis were performed under standard conditions, using 35 μg of whole-cell lysate protein per lane. In short, nitrocellulose membranes were blocked with 3% milk/PBS and incubated with the primary antibody (1:200-1:1000) overnight at 4°C. The membranes were washed thrice for 10 min in 3% milk/PBS and then were incubated with an HRP-coupled secondary antibody (1:1000) for 1 h at room temperature. The membranes were washed and the bands were visualized on x-ray film (Agfa, Cologne, Germany) using the enhanced chemiluminescent (ECL) method (Amersham Biosciences, Buckinghamshire, United Kingdom).
Cell cycle analysis. The HNSCC cells were grown as described above. The supernatants containing dead cells were pooled together with the trypsinized cells from the culture dish. After fixation in ethanol they were further incubated for 1 h in the presence of 50 μg/ml RNase A (Sigma, Taufkirchen, Germany). The suspension was applied to a 50 μm pore sized filter (Filcons, DakoCytomation GmbH, Hamburg, Germany) to remove larger cell aggregates. The cells were finally stained with 125 μg/ml propidium iodide (Sigma, Taufkirchen) and FACS analysis was performed with a BD FACSCalibur Flow Cytometer (Becton Dickinson, Heidelberg, Germany). Calculation of the data was performed with ModFit LT™ software for Mac systems (Verity Software House, Topsham, ME, USA).
Results
Sequence analysis. The RT-PCR sequence analysis demonstrated the presence of the R497K polymorphism in 4 out of the 11 (36%) cell lines. No EGFR mutations were found in any of the tested cell lines. Silent mutations, not resulting in amino acid changes, were found in 8 out of the 11 (73%) cell lines (Table I). Five out of the 7 (71%) HNSCC cell lines with an arginine (R) at position 497 of the EGFR (EGFRR497) but only 1 out of the 4 (25%) cell lines harbouring a lysine (K) at the same position (EGFRK497) were derived from tumors that presented lymph node metastases (N) at the time of diagnosis (Table I).
Influence of AG1478 treatment on cell cycle distribution and EGFR activation. Treatment of the UM-SCC-4, -14A, -27, UMB-SCC-745, -864, -969, and UT-SCC-24A, -26A cells with the EGFR-specific kinase inhibitor AG1478 resulted in a shift of cells into the subG1-phase (dead cells) (Figure 1a). Also, a rise (UM-SCC-1, -3, UMB-SCC-969, UT-SCC-26A) or reduction (UM-SCC-27, UMB-SCC-745, UT-SCC-24A) of cells in G0/G1, as well as a rise (UM-SCC-27) or reduction (UM-SCC-1, -4, -22B, -14A, UMB-SCC-745, - 969) of cells in G2/M was noted after treatment with the inhibitor. Except for UM-SCC-4 and -27, all the cell lines reacted with a reduction of cells in the S-phase of the cell cycle. The EGFRK497 cell lines included on average more cells in the subG1 and G2/M-phases than the EGFRR497 cells. However, this difference did not reach significance particularly since one of the R497 cell lines, UMB-SCC-745, in sharp contrast to the other R497 lines, most strongly reacted to AG1478, resulting in nearly nine times as many cells in the subG1-phase compared to the controls. No marked differences between the R497 and K497 groups were seen in the G0/G1-, S- and G2/M-phases (notice the scale differences in the subG1-phase compared with G0/G1-, S- and G2/M-phases). The corresponding Western blot analysis revealed a reduction of EGFR phosphorylation (Figure 1b, EGFR-p) after treatment with AG1478. This effect appeared to be more efficient in the K497 cell lines, whereas the R497 cells still presented substantial p-EGFR expression compared to the initial (untreated) values. The phosphorylation-inhibition of Erk-1/2 (p-Erk-1/2) was more efficient in the K497 cells, particularly visible in the UM-SCC-14A, -27 and UMB-SCC-969 cells, whereas the R497 cells UM-SCC-1, - 3, UMB-SCC-864 and UT-SCC-26A did not show a reduction of p-Erk-1/2 after exposure to AG1478 at all and only a slight effect was seen in the UM-SCC-4 cells. Only the UM-SCC-22B and particularly the UMB-SCC-745 cells, the latter being the very same line that responded with a dramatic rise of subG1 cells, showed a clear reduction of Erk-1/2 phosphorylation.
Clinical data of HNSCC cell lines used in the study and corresponding result of EGFR sequencing.
Influence of cetuximab treatment on cell cycle distribution and EGFR activation. As observed after the AG1478 treatment, except for the UM-SCC-1, -22B and UMB-SCC-745 cells, exposure to cetuximab resulted in a pronounced shift of cells into the subG1-phase in all the other tested cell lines (Figure 2a) although it was less pronounced than in AG1478-treated cells. A slight rise (120%) of cells in G0/G1 was noted in the UM-SCC-1 cells whereas all the other cell lines did not exhibit any notable changes of this cell population. In the UM-SCC-27 and UT-SCC-26A cells the proportion in G2/M dropped to about 60% of the control values whereas the cell lines UM-SCC-1, -22B, UMB-SCC-745 and -969 barely decreased more than 10%. A reduction of cells in S-phase was noted in all the cell lines, except for UM-SCC-4 and UMB-SCC-745, dropping to around 50% in the UM-SCC-1 cells. The EGFRK497 cell lines, as observed for the AG1478 treated cells, included on average more cells in subG1-phase than the EGFRR497 cell lines. No differences between these two groups were seen when comparing the proportions of G0/G1-, G2/M- and S-phase cells. In sharp contrast to the treatment with AG1478, the Western blot analysis revealed a seemingly paradoxical hyperphosphorylation of EGFR, particularly visible in the UM-SCC-1, -3, UMB-SCC-745, -864 and UT-SCC-24A cells, but did not influence its downstream inhibiting impact on p-Erk-1/2 (Figure 2b, p-EGFR, p-Erk-1/2). An association of Erk-1/2 phosphorylation levels with changes in the cell cycle was less obvious than after the AG1478 treatment, although in the K497 cell lines, the observed reduction of phospho-Erk-1/2 in the UM-SCC-27 and UMB-SCC-969 cells seemed to correlate well with the level of dead cells in the subG1 phase (Figure 2a and b).
Discussion
Most reports have found that activating mutations in the kinase domain of EGFR, as observed in NSCLC, are very infrequent in HNSCC tumors (17, 18). Although single studies have reported a frequency of about 7.3% (3 out of 41 HNSCC tumors) (19) mutations in EGFR indeed appear to be a rare event in HNSCC cells. Consistent with these observations none of the eleven tested HNSCC cell lines in the present study had mutations in any part of the EGFR coding sequence other than silent mutations that did not result in amino acid changes. Interestingly, in four of the cell lines, the codon 497 (alternative terminology: codon 521) polymorphism (AGG>AAG) was detected, resulting in a change of arginine (R) to lysine (K) at this position (R497K).
Bandres et al. observed an association of chemoradiotherapy response and EGFRK497 polymorphism in HNSCC tumors (5). Similarly Klinghammer et al. found a correlation of skin toxicity as well as outcome of HNSCC patients after treatment with cetuximab and docetaxel with the EGFRK497 genotype (10). Although only a limited number of HNSCC cell lines were investigated in the present study, remarkably, 71% of the EGFRR497-expressing cell lines were derived from tumors that had already metastasized into regional lymph nodes (N+) at the time of diagnosis, whereas only one of the EGFRK497 expressing cell lines (25%) was N+. Also, following treatment with the EGFR-specific kinase inhibitor AG1478, a strikingly higher number of EGFRK497-positive cells died and were found in the subG1 phase of the cell cycle compared to the EGFRR497-positive cells. A similar observation was made after treating the cells with cetuximab.
Analysis of cell cycle distribution and EGFR signaling in HNSCC cell lines treated with the EGFR-specific kinase inhibitor AG1478. a: Cell cycle phase ratio (%) of treated versus untreated tumor cells. ▪ Cell lines expressing R497, 
cell lines expressing K497. b: Western blot analysis of the corresponding cell lysates. HNSCC cell lines expressing EGFRK497 are shown on the right and those expressing EGFRR497 on the left. Actin was used as an internal control.
Analysis of cell cycle distribution and EGFR signaling in HNSCC cells after treatment with the monoclonal anti-EGFR antibody cetuximab (Erbitux®). a: Cell cycle phase ratio (%) of treated versus untreated tumor cells. ▪ Cell lines expressing R497, 
cell lines expressing K497. b: Western blot analysis of the corresponding cell lysates was performed to evaluate changes in the EGFR signaling molecules EGFR, p-EGFR, Erk-1/2 and p-Erk-1/2. HNSCC cell lines expressing EGFRK497 are shown on the right and those expressing EGFRR497 on the left. The correlation of reduction in phospho-Erk-1/2 and rise of subG1 cells was most obvious in EGFRK497 cells (see text for details). Actin was used as an internal control.
The MAP (mitogen-activated protein) kinases Erk-1 and - 2 are downstream intermediate signaling molecules in the EGFR signaling pathway. An effective inhibition of EGFR activation should therefore also be accompanied by a reduction of Erk-1/2 activation, reflected in a reduction of phospho-Erk-1/2. Changes in the levels of phospho-EGFR after inhibitor treatment, considering phospho-EGFR to be the active form of the receptor, does not always reliably predict the consequences for downstream signaling events. In a previous study, we showed that the treatment of HNSCC cells with cetuximab resulted in a paradoxical hyperphosphorylation of EGFR, likely due to a possible ligand-mimicking effect of this antibody. However, cetuximab still inhibited the downstream phosphorylation of Erk-1/2. It could be that the autophosphorylated supposedly active forms of EGFR were sequestered to a subcellular compartment and were not able to emit downstream signals or that signaling was inhibited by other EGFR-interacting molecules (16). In the present study, similar paradoxical phosphorylation of EGFR was also observed. It therefore appears of utmost importance to regularly include downstream signaling molecules such as Erk-1/2 to more reliably assess the effects of anti-EGFR targeted therapies on downstream signaling events. In most of the tested HNSCC cell lines where a reduction of phospho-Erk-1/2 expression was demonstrated, a rise of dead cells in the subG1-phase of the cell cycle also occured. Furthermore cells harbouring the EGFRK497 polymorphism appeared to be more sensitive to inhibitor treatment as shown by a more effective reduction of phospho-Erk-1/2.
Taken together, it can be stated that EGFR mutations in HNSCC cell lines appear to be a rare event. The R497K EGFR polymorphism on the other hand is found in a substantial number of cell lines. EGFRK497 carrying cell lines are more likely to be susceptible to treatment with the kinase inhibitor AG1478, as well as the monoclonal antibody cetuximab, compared to EGFRR497-positive cells. The response to inhibitors is associated with effective suppression of phospho-Erk-1/2. These results therefore support previous reports pointing to the R497K polymorphism as a positive predictor for therapy response in HNSCC patients.
Acknowledgements
The technical assistance of Ms. R. Peldszus and Ms. G. Sadowski (Department of Otorhinolaryngology, Head and Neck Surgery, University Hospital Giessen and Marburg, Campus Marburg, Marburg, Germany) and Ms. M. Alt, Dr. S. Müller-Brüsselbach and Dr. W. Meissner (Institute for Molecular Biology and Tumor Research, Philipps University, Marburg, Germany) is greatly appreciated. This study was partly funded by the Alfred und Ursula Kulemann Stiftung, Philipps University, Marburg, Germany.
- Received September 7, 2010.
- Revision received December 2, 2010.
- Accepted December 3, 2010.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
