Cancer Research Meeting Calendar Advances in Breast Cancer Research
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kong, A.
Right arrow Articles by Larijani, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kong, A.
Right arrow Articles by Larijani, B.
[Cancer Research 66, 2834-2843, March 1, 2006]
© 2006 American Association for Cancer Research

Clinical Research

Prognostic Value of an Activation State Marker for Epidermal Growth Factor Receptor in Tissue Microarrays of Head and Neck Cancer

Anthony Kong1,2, Pierre Leboucher4, Russell Leek3, Véronique Calleja1, Stuart Winter3, Adrian Harris3, Peter J. Parker2 and Banafshé Larijani1

1 Cell Biophysics Lab, 2 Protein Phosphorylation Lab, London Research Institute, Cancer Research UK, London, United Kingdom; 3 Weatherall Institute of Molecular Medicine, Cancer Research UK, University of Oxford, John Radcliffe Hospital, Headley Way, Oxford, United Kingdom; and 4 Laboratoire de Physiologie de la Pérception et de l'Action, College de France, Paris, France

Requests for reprints: Banafshé Larijani, Cell Biophysics, 44 Lincoln's Inn Fields, Cancer Research UK, London, WC2A 3PX, United Kingdom. Phone: 44-20-7269-3082; Fax: 44-20-7269-3094; E-mail: banafshe.larijani{at}cancer.org.uk.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Overexpression and mutation of epidermal growth factor receptors (EGFR) have been shown to be important in the prognosis of several cancers, including head and neck cancers. However, our inability to define the activation status of these and other receptors limits our ability to assess the importance of these pathways and to exploit effectively new molecularly targeted treatments directed at their catalytic activities. Here we describe the use of automated, high-throughput fluorescence lifetime imaging microscopy to measure EGFR autophosphorylation status by fluorescence resonance energy transfer (FRET) in head and neck tumors. We have correlated FRET efficiency with the clinical and survival data. The results from head and neck arrays show that high FRET efficiency is correlated with worsening disease-free survival but not with overall survival. This powerful tool could be exploited as a new independent quantitative prognostic factor in clinical decisions and cancer management. (Cancer Res 2006; 66(5): 2834-43)


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer in the world and represents 6% of all cancers (1, 2). It comprises a heterogeneous group of tumors in terms of its etiology and molecular mechanisms (3). Traditionally, the management of this cancer represents a combination of surgery, radiotherapy, and chemotherapy (4). However, molecular therapy is gaining importance in this group of cancers. Elevated levels of transforming growth factor {alpha} (TGF-{alpha}) and epidermal growth factor receptor (EGFR) mRNAs have been detected in the normal mucosa samples of head and neck cancer patients, and hence implicated in the pathogenesis of head and neck cancer (5), and have been shown to be significant predictors of disease-free survival in these patients (6). In addition, several studies have also shown correlation of overexpression of EGFR with overall survival, disease-free survival, increased locoregional recurrence, and decreased sensitivity to radiation treatment in these patients (7–10). To assess EGFR overexpression, its expression determined by immunohistochemistry has been the standard. However, not all studies have found a correlation of overexpression of the EGFR and prognosis (11, 12). Furthermore, the EGFR level does not necessarily predict the response to molecular therapy (13). Part of the variation may derive from the fact that immunohistochemistry does not provide a truly quantitative analysis of EGFR expression. The results tend to vary between the laboratories because of the lack of a standardized scoring system and the subjective process of the interpretation of stained samples (14). Moreover, fixatives used and the storage time of unstained tissue sections are the cause of variations (15). In a recent article on molecular classification of HNSCC using patterns of gene expression, it was found that the poorest outcome group was the group with tumors characterized by the high expression of TGF-{alpha} with evidence of activation of EGFR pathway (16). In addition, one study has found that it is the overexpression of multiple receptors, mainly EGFR1 (ErbB1), with other ErbB receptors (ErbB2-4) that correlates more with metastatic disease (17). Based on our understanding of the behavior of these receptors, these studies indicate that not only is the actual level of EGFR expression important, but more so the activation/phosphorylation state of these receptors.

The current methods of measuring EGFR levels, including immunohistochemistry, cannot be unequivocally endorsed as predictive of patient prognosis or response to treatment and they are also unable to detect the functional status of the EGFR pathway (18). Therefore, there is a need for a quantitative method to be set up to measure EGFR and its activation status. To achieve this, we have exploited fluorescence resonance energy transfer (FRET) monitored by fluorescence lifetime imaging microscopy (FLIM) as a molecular prognostic tool to identify HNSCC patients who show overexpression and/or phosphorylation of the EGFR. Using A431 cells as a test bed, we have evaluated the use of a suitable pair of donor (Cy3b) and acceptor fluorophores (Cy5) conjugated to F4 (EGFR cytoplasmic domain antibody) and FB2 (antiphosphotyrosine), respectively, and monitored FRET between the fluorophores detected by FLIM on EGF stimulation. We have illustrated that the FRET associated with the coincident binding of the labeled monoclonals is specific. These validated reagents have been applied to head and neck tumor arrays using for the first time a high-throughput automated FLIM. Employing this highly selective FRET assay in tumor arrays, we show that increased FRET can be correlated with disease recurrence and prognosis of the patients. Moreover, we have investigated whether the EGFR status determined by immunohistochemistry or by average fluorescence intensity is also correlated with activation of EGFR.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Patients. The work was carried out with the approval of the Oxford Ethics Committee. One hundred thirty consecutive cases of HNSCC from the Ear Nose Throat Department, Radcliffe Infirmary (Oxford, United Kingdom) were analyzed. Surgical approach was always the primary approach in these patients unless they were not encompassable in an appropriate surgical field. If, after primary surgery, margins were positive, then radiotherapy was given to the patients. No chemotherapy was given to these patients. No primary radiotherapy cases are included here and patients who had previously been treated for HNSCC were also excluded.

Materials and cell lines. A431 cells were obtained from cell services at Cancer Research UK London Research Institute. F4 immunoglobulin G1 (IgG1) mouse antibody, residues 985-996 (monoclonal against EGFR cytoplasmic domain), and FB2 IgG3 (monoclonal against phosphotyrosine) antibodies were obtained from the Monoclonal Antibody Lab, London Research Institute, Cancer Research UK. The secondary sheep anti-mouse IgG was purchased from Amersham Biosciences UK Ltd. (Little Chalfont, Buckinghamshire, United Kingdom). AG 1478, a selective inhibitor of the EGFR tyrosine kinase (IC50 = 3 nmol/L), was from Calbiochem UK (Nottingham, United Kingdom). The monoconjugated fluorophores Cy3B and Cy5 were from Amersham Biosciences. Protein tyrosine phosphatase from Yersinia enterocolitica (recombinant, E. coli) was purchased from Calbiochem.

Western blotting. A431 cells were grown to 80% confluency in six 10-cm cell culture plates. Four of the six plates of cells were treated with different doses of AG 1478 (0.3, 0.6, 1.5, and 3 µmol/L) for 2 hours. All six plates apart from one were then stimulated with EGF (50 ng/mL) for 10 minutes after being serum starved for 16 hours. The cells from all the plates were then lysed in lysis buffer (20 mmol/L Tris-HCl, 150 mmol/L NaCl, 100 mmol/L NaF, 10 mmol/L Na4P207, 10 mmol/L EDTA with 1% Triton and protease inhibitor cocktail; Roche) and centrifuged at 4°C with the removal of the cell pellets. PAGE was carried out employing 10 µg of protein in each lane. Western blots were done using F4 and FB2 antibodies. A dilution of 1:15,000 of F4 (stock concentration 1.7 mg/mL) and a dilution of 1:500 of FB2 (stock concentration of 1.3 mg/mL) were used to probe the cell lysates. Antibodies were incubated overnight at 4°C. They were detected using horseradish peroxidase–linked secondary antibody (a dilution of 1:5,000 sheep anti-mouse IgG) and visualized with an enhanced chemiluminescent system (Amersham).

Fluorescence resonance energy transfer by fluorescence lifetime imaging microscopy. FRET involves the transfer of energy from an excited donor molecule to a nearby (<7 nm) spectrally overlapping acceptor. FRET can be quantified by measuring fluorescence lifetime of the donor, which is reduced as energy is nonradiatively transferred via a dipole-dipole interaction. Spatial aspects of fluorescence lifetime may be assessed by using FLIM (19). In this study, we have monitored donor lifetime variations in the frequency (phase) domain where the excitation light is sinusoidally modulated at 80.218 MHz to excite the sample. The emitted light oscillates at the same modulation frequency but with a phase shift and a decrease in amplitude (demodulation). Determining these two variables permits measurement of phase ({tau}p) and modulation depth ({tau}m) of the fluorescence. The lifetime, <{tau}>, is the average of phase shift and relative modulation depth, ({tau}m + {tau}p) / 2, of the emitted fluorescence signal. The efficiency of FRET can be mapped in a single cell by determining the excited state lifetime of the donor fluorophore at each pixel of the image. FRET efficiency is calculated as E = 1 – [<{tau}d/a> / <{tau}d>] x 100, where <{tau}d/a> is the average lifetime of the donor in the presence of the acceptor and <{tau}d> is the average lifetime of the donor in the absence of the acceptor (19, 20).

Conjugation of donor and acceptor fluorophores to antibodies. F4 was conjugated to Cy3b (donor fluorophore) and FB2 to Cy5 (acceptor fluorophore). 100 µL of N,N-dimethylformamide was added to 1 mg Cy3b to make a 10 mg/mL stock solution (15 mmol/L). The stock 10 mg/mL Cy3b was diluted in N,N-dimethylformamide by 10-fold to 1 mg/mL (1.5 mmol/L). Fifty microliters of Cy3b/N,N-dimethylformamide from a stock of 1 mg/mL were added dropwise into 450 µL F4:50 µL Bicine (1 mol/L, pH 8) with continuous stirring. The final concentration of F4-Cy3b was ~100 µg/mL (150 µmol/L). The solution was stirred in the dark for 1 to 2 hours. To conjugate FB2 with Cy5, 20 µL of N,N-dimethylformamide were added to a Cy5 vial. FB2 dye in N,N-dimethylformamide was then added dropwise to 450 µL FB2:50 µL Bicine (1 mol/L, pH 8) while stirring. The solution was stirred in the dark for 1 to 2 hours. Then the conjugated antibodies were separated from free dyes by column chromatography. The dye/protein ratios were maintained constant per experiment. The dye/protein ratios were 1.6 for F4-Cy3b and 2.3 for FB2-Cy5. The dye/protein ratios were measured by UV/Visible spectroscopy at 280 nm to determine F4 and FB2 concentrations. The concentrations of F4-Cy3b and FB2-Cy5 were detected at 516, 552, and 650 nm, respectively. The dye/protein (D/P) ratios were calculated using the protocol provided by Amersham Biosciences for Cy3B monoreactive dye:

Formula

FRET Experiments on A431 cells. A431 cells were grown in 24-well plates with coverslips after seeding ~15,000 per well. The cells were then stimulated with EGF 50 ng/mL for 10 minutes following serum starvation for 16 hours. Following stimulation, the cells were fixed with 4% paraformaldehyde at room temperature for 10 minutes. Five hundred microliters of 0.2% (v/v) Triton X-100 were added in the well for 5 minutes to make the cell membrane permeable followed by 1 mg/mL fresh sodium borohydrate/PBS for 10 minutes to quench background fluorescence. The cells were then blocked with 1% w/v bovine serum albumin (BSA)/PBS for 30 minutes. To label the cells with conjugated F4-Cy3b, we incubated the cells with 10 µg/mL F4-Cy3b/1% BSA/PBS for 1 hour. For cells that required detection with the acceptor fluorophore, we incubated them further with conjugated phospho-EGFR antibody, FB2-Cy5, for 1 hour. The coverslips were mounted on the slide with Mowiol mounting medium containing 2.5% (w/v) 1,4-diazabicyclo (2,2,2) octane as an antifade. The slides were left at 37°C for 1 hour and then at room temperature overnight before image acquisition.

For FRET experiments, all images were taken using a Zeiss Plan-Apochromat x100/1.4 NA phase 3-oil objective with images recorded at a modulation frequency of 80.218 MHz. The donor (F4-Cy3b) was excited using 514-nm line of an argon/krypton laser and the resultant fluorescence was separated using a combination of dichroic beam splitter (Q565 LP, Chroma Technology Corp., Rockingham, VT) and narrow band emitter filter (BP 610 /75; Chroma Technology Corp.). The two-dimensional histograms show both the phase ({tau}p) and modulation ({tau}m) measurements for lifetime for all recorded pixels and these quantify the reduction of the phase and modulation lifetimes of donor Cy3b when there is FRET. A series of these experiments were done and optimum concentrations of AG 1478 and EGF, dilution factor of FB2-Cy5, and incubation periods of FB2-Cy5 were determined before applying the cell variables to tissue microarrays. For those experiment involving the protein tyrosine phosphatase from Y. enterocolitica (YOP), 50 units of phosphatase in 50 µL reaction buffer [50 mmol/L Tris-HCl (pH 7.2), 150 mmol/L NaCl, 5 mmol/L DTT, 2.5 mmol/L Na2EDTA, and 100 µg/mL BSA, as recommended by Calbiochem] were used for each coverslip on the laboratory film (parafilm "M").

Tissue microarrays—lifetime and average intensity measurements. Tissue arrays were prepared from formalin-fixed, paraffin-embedded tumor blocks derived by surgical resection from the Pathology Department, John Radcliffe Hospital (Oxford, United Kingdom). Quality control for tumor specimens has been undertaken. All slides were examined by a specialist consultant pathologist. It is not possible for each core to be uniformly involved by the tumor and naturally there is variation between the individual cores with the percentage of the tumor. However, it was not thought appropriate to correct for the percentage of the area occupied by the tumor because EGFR was only expressed in the tumor and not in stromal cells by our immunohistochemistry studies and the analysis undertaken here measures the intensity of activation of the receptor. Serial sections adjacent to those used in the FRET analysis were stained with H&E to confirm the presence of tumor in the cores. The tonsil sections are provided as a reference point for slide orientation and contain heterogeneous tissues including squamous, lymphoid, and vascular tissues as well as supporting stroma. Sections which contained squamous tissue were analyzed in the study. Cores of 500 µm were taken and transferred to an 8 x 15 core recipient array block using a Beecher Instruments Manual Tissue Arrayer-1 (Beecher Instruments, Inc., Sun Prairie, WI). Five-micrometer sections were cut from each array block and dewaxed in xylene. The sections were then rehydrated through graded alcohols to water.

For antigen retrieval process, 0.01 mol/L citrate buffer (pH 6) was heated in an 800-W microwave oven at full power for 4 minutes. The sections were then placed in the buffer and heated in the microwave for a further 4 minutes at full power. The citrate buffer was then topped up and microwave heated again for 4 minutes. The citrate solution and sections were then allowed to cool for 10 minutes before handling.

To label the tumor arrays with F4-Cy3b, we immersed the arrays with 400 µL of 0.2% (v/v) Triton X-100 for 5 minutes to permeabilize the membrane, followed by 1 mg/mL fresh sodium borohydrate/PBS for 10 minutes to quench background fluorescence. 1% (w/v) BSA in PBS was used for blocking. The conjugated EGFR antibody F4-Cy3b, stock solution 100 µg/mL, was diluted in 1% BSA/PBS to 10 µg/mL. We incubated the tumor arrays with 400 µL of 10 µg/mL F4-Cy3b/1% BSA/PBS on the slide for 1 hour and then fixed the arrays with 4% paraformaldehyde for 10 minutes. To label the tumor arrays with the acceptor, we incubated the tumor slide with FB2-Cy5 (diluted 2-fold) for 1 hour. The arrays were then mounted on a glass cover with Mowiol mounting medium containing 2.5% (w/v) 1,4-diazabicyclo (2,2,2) octane.

Using the automated FLIM system, we mapped each tumor core according to the position on the arrays. Images from each tumor core were acquired automatically from the arrays according to their positioning. The phase, modulation, and average lifetimes of each tumor core were calculated automatically and the average FRET efficiency for each tumor core was obtained. Average FRET efficiency for each tumor core was correlated with survival data. We first carried out a pilot study to assess the first set of arrays with the automated system. To validate the pilot study, we prepared a new set of arrays with new antibodies and fluorophores and processed the arrays with automated FLIM. In the validation study, the system was programmed to run two loops so that two lifetime measurements were taken from each tumor core (n = 2; i.e., 572 tumor cores in total). However, in any one array a patient would have a duplicate sample; thus, for each patient the final lifetime represents an average of four measurements. A total of 1,114 tumor core recordings were made and analyzed in the validation study.

To obtain the EGFR status by FLIM, we calculated the donor intensity. For each tumor core, images were acquired and an intensity distribution histogram was plotted using Matlab version 7 and the median of the intensity distribution was calculated by Matlab 7. This median was normalized to the background intensity. We used the average of two median intensity values from the two tumor cores of the same patient in each array. Thereafter, these values were used to correlate with immunohistochemistry stains and average FRET efficiency.

EGFR immununohistochemistry staining. Five-micrometer sections were dewaxed in xylene and rehydrated through graded alcohols to water. High-temperature antigen retrieval was done in 0.01 mol/L citrate buffer (pH 6) heated for 1 minute in a microwaveable pressure cooker. The sections were then cooled and washed in water. Following peroxidase block, sections were protein blocked for 15 minutes in 2.5% normal horse serum. The primary monoclonal antibody F4 was incubated on the sections for 1 hour at room temperature at a concentration of 10 µg/mL and then washed twice in PBS. The primary antibody was visualized using a horseradish peroxidase Vector Labs Immpress Mouse Kit (Vector Laboratories, Burlingame, CA) and the sections were developed with diaminobenzidine. After washing in water, sections were mounted on coverslips in aqueous mounting medium.

Statistical analysis. Disease-free survival is defined as the length of time after treatment during which no cancer is found, and overall survival is a defined period of time that the subjects in a study have survived since diagnosis or treatment. All statistical analyses were done using GraphPad Prisms 3cx (Macintosh Version) except the univariate and multivariate analyses for prognostic factors which were done using "R" (see below). The Kaplan-Meier survival curves were used to compare between the groups and a log-rank test was used to assess the hazard ratios. For the univariate and multivariate analysis of the prognostic factors, we analyzed the data by fitting Cox proportional hazard models to the data with top 10% FRET efficiency as the group indicator with a common baseline hazard function, and including immunohistochemistry, Unio Internationale Contra Cancrum (UICC) tumor stage, grade, necrosis, age, and sex as covariates. To fit the models, we used the "coxph" function from the "survival" package within R. R is a language and environment for statistical computing and is available as Free Software under the terms of the Free Software Foundation's GNU General Public License in source code form.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Detection of EGFR phosphorylation in A431 cells by FRET. Traditionally immunochemistry has been used to measure the level of EGFR expression in tumors without informing on activation status. To develop an activation state readout with suitable specificity for the EGFR, we postulated that following EGFR activation, FRET would be detectable in fixed cell samples between F4-Cy3b (a monoclonal to the cytoplasmic domain of the receptor linked to a donor fluorophore, Cy3b) and FB2-Cy5 (an EGFR autophosphorylation site monoclonal linked to an acceptor fluorophor, Cy5). A431 was chosen as our test cell line to validate the assay because of its high EGF receptor expression and its extensive use for nearly 20 years in the analysis of EGFR function. Initially, the specificity of the antibodies F4 and FB2 was tested through Western blot of A431 cell lysates (Fig. 1A). The receptor is specifically detected by the F4 monoclonal whereas FB2 recognizes both the phosphorylated receptor and two pairs of faster migrating (phospho-) proteins. The degree of immunorecognition of the receptor by FB2 specifically increased in response to EGF as expected for this phosphotyrosine site-directed monoclonal. Importantly, this immunoreactivity was reduced in a dose-dependent manner by the selective EGFR tyrosine kinase inhibitor AG 1478. It is notable that in these cells, there is a basal degree of EGFR phosphorylation consistent with prior data indicating a degree of autocrine receptor activation (21). It is also important to note that the additional recognition of faster migrating protein species by the FB2 monoclonal does not interfere with the two-site FRET assay reported here because the specificity of the analysis is determined by F4 (this is an important feature of the two-site immunohistochemistry assay; see further below).


Figure 1
View larger version (34K):
[in this window]
[in a new window]
  		Figure 1. Autophosphorylation of the EGFR is prevented by AG 1478 on EGF stimulation. Its activation is monitored by FRET. A, A431 cells were stimulated with 50 ng/mL of EGF for 10 minutes. Cell lysates were probed with F4 (antibody against the cytoplasmic domain of EGFR) and FB2 (EGFR autophosphorylation site monoclonal). The 170-kDa F4 immunoreactive band (EGFR) comigrates with the major signal from FB2 (arrow). Cells were pretreated with increasing doses of AG 1478 as indicated and then stimulated with EGF. Part of the basal phosphorylation seen in A431 cells without EGF stimulation is likely to be due to autocrine loop activation. The phosphorylation increased with EGF stimulation. AG 1478 (0.3 µmol/L), a tyrosine kinase inhibitor, was inadequate for inhibiting autophosphorylation; however, higher doses of the inhibitor abolished the autophosphorylation. B, the two-dimensional histogram of phase and modulation lifetimes shows that the lifetimes of F4-Cy3b alone do not change in the presence of growth factor EGF (n = 3). However, the lifetimes decrease (diagonal decrease) in the presence of acceptor (FB2-Cy5) even without EGF (n = 3). The lifetime values are indicated in the table. The lifetimes decrease further in the presence of EGF (n = 3), indicating here that this two-site FRET assay can be used to report EGFR activity. With AG 1478 at 1.5 and 3 µmol/L, the lifetimes of Cy3b return to higher lifetimes (n = 6). C, the lifetime maps of A431 cells show an average lifetime for F4-Cy3b of 1.63 ns. The average lifetime remains the same in the presence of EGF but decreases with the acceptor (FB2-Cy5). The decrease in average lifetime to 1.38 ns following stimulation with EGF is most noticeable at the plasma membrane as seen from the color changes of the lifetime map. In the presence of AG 1478 (1.5 and 3 µmol/L), the average lifetime returns to a higher lifetime of 1.49 ns.

 
Fixed A431 cells were employed to test whether, in an immunohistochemistry format, the coincident binding of F4-Cy3b and FB2-Cy5 to phosphorylated EGFR produced a specific FRET signal. Cells were treated with or without EGF, fixed, and processed as described in Materials and Methods. Employing this assay, it was found that following EGF stimulation, there is a marked increase in FRET as illustrated in Fig. 1B. This is a specific property of the coincident binding of the two fluorescently labeled antibodies because the lifetime change observed for the donor fluorophore following EGF treatment is not observed if the antibody-acceptor conjugate is omitted. In the absence of the EGF stimulus, there is a reduced degree of FRET but with lifetimes for the donor in the presence of the acceptor below those observed for donor alone (see Fig. 1B).

To determine whether the EGF-induced increase in FRET reflected an increase in immunorecognition of tyrosine-phosphorylated receptor by the FB2 antibody-acceptor conjugate, cells were pretreated with AG 1478. At effective doses of this EGFR inhibitor (1.5 and 3.0 µmol/L; see Fig. 1A), FRET was reduced to the basal unstimulated level (Fig. 1B). AG 1478 treatment of the MDAMB-231 cell line produced the same pattern of results as A431 cells (Supplementary Fig. S1A; Supplementary Table S1A). Dose dependency studies in A431 cells showed that at lower concentrations (0.3 and 0.6 µmol/L), FRET was not reversed (Supplementary Table S1B). Pretreating the cells with AG 1478 did not reverse the basal degree of FRET. To verify that the signal was indeed due to phosphotyrosine recognition, we employed the protein tyrosine phosphatase from Y. enterocolitica (YOP). Pretreatment with YOP abolished both the basal activity and EGF-induced FRET (Supplementary Fig. S1B; Supplementary Table S1C). This suggests that FRET between F4-Cy3b and FB2-Cy5 in the basal state is not due to the direct autoactivation of the ErbB1 receptor (AG 1478 sensitive) but is probably mediated by heterodimerization with other activated EGFRs or by other receptor-associated proteins, which are not inhibited by AG 1478.

An example of donor Cy3b intensity images and their corresponding average lifetime maps from the same FRET experiment are illustrated in Fig. 1C. The average lifetime of F4-Cy3b in these studies decreased from 1.63 to 1.50 ns when the acceptor FB2-Cy5 was present and decreased further to 1.38 ns on EGF stimulation. The changes in lifetime induced by EGF are localized mainly at the plasma membrane as seen from the pseudocolor changes of the lifetime map (Fig. 1C). This is entirely consistent with the expected increase in tyrosine phosphorylated receptor at the cell surface following a 10-minute stimulation with EGF. Concentrations of 1.5 and 3 µmol/L AG 1478 reversed the donor lifetime to 1.49 ns.

The observed ligand-induced FRET, its inhibitor-dependent reduction, and the defined plasma membrane response, coupled with the specificity of the F4 monoclonal antibody, validate the use of the F4-FB2 antibody pair to monitor EGFR expression (F4 immunoreaction) and EGFR activation (F4-FB2 coincident immunorecognition) in an immunohistochemistry format.

The activation state of EGFR in head and neck tumors. We used this validated two-site FRET assay to determine the pattern of EGFR phosphorylation in a series of HNSCC. An archive of head and neck tumor samples embedded in paraffin blocks was converted into three arrays. The arrays contained 286 tumor cores (a duplicate of a tumor core from each patient) and several cores comprised normal tonsil tissue as negative controls (Fig. 2A). These arrays were prepared for FRET experiments using the variables established in A431 cells.


Figure 2
View larger version (31K):
[in this window]
[in a new window]
  		Figure 2. EGFR activity can be mapped by variations in FRET in head and neck tumor cores. A, left, diagrammatic representation of a standard Oxford 8 x 15 tissue microarray. Here we had six such arrays for each experiment: three sets containing 286 tumor cores to be labeled with F4-Cy3b alone and a duplicate of the three sets to be labeled with F4-Cy3b and FB2-Cy5. Right, an example of a tumor core with conventional EGFR immunostaining of F4 antibody (monoclonal against EGFR cytoplasmic domain, which is the same monoclonal antibody used in FRET experiments). B, left, intensity images of tumor cores from the array. Right, lifetime maps of the cores. The upper tumor core of patient 1 was from an array labeled with donor alone (F4-Cy3b). The average donor lifetime of the tumor core was 2.00 ns. The donor and acceptor (F4-Cy3b + FB2-Cy5) core was from a duplicate array. There was no change in the lifetime with a FRET efficiency of 0%. The second set of donor and acceptor cores from patient 2 indicates a further decrease of the average lifetime (2.20-1.90 ns) with a FRET efficiency of 11%. The decrease in donor lifetime induced by FRET is used to calculate the FRET efficiency (see Materials and Methods).

 
A duplicate of each array was also prepared so that one array was labeled with donor alone (F4-Cy3b) and the other with donor and acceptor (F4-Cy3b + FB2-Cy5). For each tumor core, we obtained a pair of average lifetimes, one from the array with donor alone and one from the array that was labeled with donor and acceptor (Fig. 2B). Comparing the two arrays, we were able to calculate average FRET efficiency for each tumor core (see Materials and Methods).

From the cell model, we determined that for each experimental condition the average FRET efficiency was reproducible. Therefore, average FRET efficiency was used as the main variable to correlate with the clinical data (Fig. 3). The survival data of 130 head and neck cancer patients were compared with average FRET efficiency. The average FRET efficiencies of the normal tonsil tissue samples were used as controls. The values of the negative controls ranged from 0% to 8.90% with a median average FRET efficiency of 4.86%. For the tumors, the average FRET efficiencies ranged from 0% to 14.70% and the median was 4.13%. The patients were ranked according to their average FRET efficiency and the groups split for comparison into (a) upper median versus lower median, (b) upper tertile versus lower two tertiles, and (c) upper 10% versus lower 90% range of average FRET efficiency. The Kaplan-Meier method was used to compare the survival curves between the groups of patients and the log-rank test was used to calculate the hazard ratios (see Materials and Methods). We found that higher FRET efficiency was correlated with worse disease-free survival and this was statistically significant in the groups split by upper tertile and upper 10% but not the median (Fig. 3A and B). Although high FRET efficiency was correlated with disease-free survival, it was not correlated with overall survival in any of the groups (Supplementary Fig. S2A and B).


Figure 3
View larger version (21K):
[in this window]
[in a new window]
  		Figure 3. Kaplan-Meier curves using average FRET efficiency as a prognostic marker illustrate that FRET efficiency is correlated with disease-free survival. A, pilot study: disease-free survival between patients in upper tertile versus the lower two tertiles of average FRET efficiency. Comparing the patients in the lower two tertiles with those in the upper tertile, the log-rank test revealed P = 0.05 and a hazard ratio of 0.57 (95% CI, 0.29-0.99). B, pilot study: disease-free survival between patients in upper 10% versus lower 90% of average FRET efficiency. Patients in the lower 90% range of average FRET efficiency had a hazard ratio of 0.43 (95% CI, 0.09-0.89; P = 0.03) compared with the patients in the upper 10% range. C, validation study: disease-free survival curves comparing the upper 10% range with the lower 90% of average FRET efficiency (P = 0.04; hazard ratio, 0.48; 95% CI, 0.14-0.95). D, the linear regression shows no correlation between donor intensity Cy3b and average FRET efficiency (r2=0). In these studies, expression of EGFR is not correlated with its phosphorylation state.

 
To ensure interassay validity, the above experiments were repeated in a new set of six tumor arrays, which contained tumor cores from the same patients. New preparations of conjugated antibodies were used and average FRET efficiencies were again calculated from each tumor core using automated high-throughput FLIM. The automated system was programmed to do multiple loops and we did two loops for each array. Thus, we acquired two measurements for each process (572 measurements were acquired). The Kaplan-Meier survival curves (Fig. 3C) were used to compare between the patients in the upper 10% FRET efficiency and those in the lower 90% efficiency and the log-rank test was used to compare the hazard ratio. The results were similar to the initial study in that the patients in the lower 90% range of average FRET efficiency had a better disease-free survival compared with the upper 10% range (Fig. 3C). However, there was no statistical significance between the upper tertile and lower two tertile groups in this validation study [hazard ratio, 0.77; 95% confidence interval (95% CI), 0.76-1.9; P = 0.35]. From these FRET efficiency studies, we deduce that phosphorylation of EGFR assessed by high-throughput FLIM correlates with disease-free survival.

Using Mann-Whitney tests, it was found that there was no evidence to suggest that average FRET efficiency was associated with grade or UICC tumor stage (Supplementary Fig. S3A and B). To address whether EGFR expression itself was correlated with its activation, we used average donor fluorescence intensity measurements (i.e., F4 immunoreactivity) and compared them with average FRET efficiency. It is important to note that average fluorescence intensity measurements of F4-Cy3b are related to the amount of receptor in each tumor core (see Materials and Methods). To test the correlation between EGFR expression and its activity, we calculated the linear regression between the average fluorescence intensity and average FRET efficiency. The analysis showed that there was no linear relationship between the two variables (r2 = 0; Fig. 3D). These studies illustrated that there was minimal correlation between EGFR concentration and its activation.

In summary, we have illustrated that there is a prominent correlation between EGFR activation and disease-free survival and a lack of correlation between EGFR expression and its activation status in HNSCC cores.

Conventional immunohistochemistry does not reveal a correlation of EGFR overexpression with disease-free survival and overall survival. Previously, immunohistochemistry analysis of EGFR expression has been shown to correlate with survival, particularly disease-free survival, and hence was considered useful in predicting disease recurrence (8, 9). To assess whether in this patient cohort conventional immunohistochemistry revealed a correlation between levels of EGFR expression and survival data, we marked the same tumors from the head and neck cancer patients with the F4 antibody using conventional methods although only 122 patient immunohistochemistry scores were obtainable due to inadequate tumor sample or loss of tumor sample from the array slide.

Table 1 shows the characteristics of patients according to subsites of head and neck tumors and Table 2 shows the tumor characteristics according to immunohistochemistry EGFR scores. The majority of the tumors were of UICC tumor stage III and IV and moderate to poor grade. Fifty-seven of the 122 patients (47%) have overexpression of EGFR scored either 2+ or 3+ with only 17 of the 122 patients (14%) scored at 3+ overexpression.


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of patients according to subsites of head and neck tumors

 

View this table:
[in this window]
[in a new window]
  		Table 2. Characteristics of tumors in relation to immunohistochemistry stains

 
As previously described, the same statistical tests were used to determine the correlation between disease-free survival, overall survival, and EGFR expression. Figure 4A and B shows that EGFR overexpression (scores 2+ and 3+) by immunohistochemistry was not correlated with disease-free survival or overall survival. Even a further division of the subgroups into EGFR 0 to 2+ versus EGFR 3+ did not show a statistically significant difference in disease-free survival (data not shown). The statistical results show that conventional immunohistochemistry does not reveal a correlation with disease-free survival and, therefore, does not seem to be informative as a single variable to predict disease recurrence in this retrospective study.


Figure 4
View larger version (20K):
[in this window]
[in a new window]
  		Figure 4. EGFR concentration does not correlate with the activity of EGFR and does not reveal a correlation with disease-free survival or overall survival. A, Kaplan-Meier curves show that there is no correlation of EGFR overexpression (2-3+) and disease-free survival (hazard ratio for EGFR (0-1+), 0.89; 95% CI, 0.5-1.6; P = 0.67, log-rank test). B, Kaplan-Meier curves also do not show a statistically significant correlation between EGFR overexpression (2-3+) and overall survival (hazard ratio for EGFR (0-1+), 0.64; 95% CI, 0.35-1.05; P = 0.74, log-rank test). C, Mann-Whitney test shows that there is no correlation between donor fluorescence intensity of F4-Cy3b with EGFR immunochemistry scoring (P = 0.39, log-rank test). D, Mann-Whitney test illustrates that there is no correlation between average FRET efficiency and EGFR immunochemistry scoring (log-rank test, P = 0.24). As in Fig. 3D, this indicates that expression of EGFR is not correlated with its activity.

 
Next, we addressed whether overall EGFR expression and EGFR phosphorylation calculated from the average fluorescence intensity of F4-Cy3b and the FRET efficiency, respectively, were correlated with EGFR expression by immunohistochemistry. To do this statistical study, the Mann-Whitney test was used. Figure 4C and D illustrates that there was no correlation between immunohistochemistry EGFR scoring and average F4 fluorescence intensity (EGFR expression) and between immunohistochemistry scoring and average FRET efficiency (P = 0.39 and P = 0.24, respectively).

Studies of EGFR expression (immunohistochemistry and average donor fluorescence intensity) when compared with activation reveal the absence of correlation between EGFR status and its activation.

Determination of the prognostic value of average FRET efficiency. Finally, we determined the prognostic value of the top 10% FRET efficiency by analysis of the disease-free survival and overall survival using univariate and multivariate analysis (see Materials and Methods). Table 3 shows the prognostic value of top 10% FRET efficiency in comparison with immunohistochemistry using univariate analysis. Top 10% FRET efficiency was a significant variable (P = 0.04) for disease-free survival but not for overall survival (P = 0.35), and overexpression of EGFR (2-3+) by immunohistochemistry was not a significant variable for either disease-free survival or overall survival. Table 4 shows the prognostic value of the top 10% FRET efficiency for disease-free survival and overall survival with UICC tumor stage and grade as covariates. The most significant variable for disease-free survival was top 10% FRET efficiency (P = 0.03) and none of the other variables were significant for overall survival by multivariate analysis. Further analysis including age and sex as covariates, in addition to UICC tumor stage and grade, showed again that top 10% FRET efficiency was the only significant variable for disease-free survival (P = 0.013; Supplementary Table S2).


View this table:
[in this window]
[in a new window]
  		Table 3. Prognostic value of the top 10% FRET efficiency by univariate analysis as compared with conventional immunohistochemistry

 

View this table:
[in this window]
[in a new window]
  		Table 4. Prognostic value of top 10% FRET efficiency by multivariate analysis with UICC tumor stage and grade as covariates

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
In this study, we have tested and validated an automated assay compatible with a microarray format that will inform on the phosphorylation status of the EGFR. The assay is designed to provide a two-site assay for phosphorylation of the EGFR using coincidence detection of protein and phospho-site selective monoclonal antibodies. This in-built specificity circumvents the requirement for monospecific modification site reagents because the specificity is determined by the protein-directed monoclonal. This is of particular importance in the context of tissue microarrays where there is no opportunity to test the validity of the immunoreactions observed. This assay has been developed for automated data capture in a format appropriate for tissue microarray analysis. We have exploited this assay with microarrays of head and neck cancers to determine the relationship between EGFR phosphorylation and prognosis. The study reveals that 10% of patients with a high degree of EGFR autophosphorylation (as evidenced by those with the highest FRET efficiency) have a poor disease-free survival. Importantly, this activation status does not correlate with EGFR concentration per se and this latter variable itself does not correlate with disease-free survival. This indicates that as might be predicted, the activation status of this receptor signaling pathway best defines its role in relation to the pattern of disease progression; as shown here, this seems to be a poor prognosis determinant with respect to disease recurrence.

Currently, the most accurate predictor of disease recurrence for head and neck cancer is the tumor-node-metastasis stage, particularly the lymph node status. There have been several studies seeking to find prognostic markers that can complement clinicopathologic information to predict survival and recurrence in HNSCC [e.g., EGFR overexpression level (8), TGF-{alpha}/EGFR mRNA (9), and p53 status (22, 23)]. At present, an elevated level of TGF-{alpha} seems to be as important or even more important in classifying prognostic groups than EGFR expression level (6, 12, 16, 24, 25). It is implicit that ligand is engaging and activating its receptor and, hence, the assessment of activation status should be informative; indeed, our data support this contention. The activation of the receptor measured by the method described here does give prognostic information that is not available from analyzing EGFR in the conventional way or even with phospho-antibodies as indicated by Chung et al. (16).

Overexpression and high levels of EGFR do not necessarily reflect the functional state of this pathway. In recent studies, immunohistochemistry reporting on EGFR expression levels has been shown to be inconsistent in its predictions of disease recurrence and notably response to treatment (14, 18). This emphasizes the need to understand at a higher level of detail the functional status of candidate markers employed in diagnostic, prognostic, and therapeutic settings. The approach exemplified here clearly can be employed in this manner and indeed can be exploited for other pathways where fixation-insensitive changes occur. Initial studies on protein kinase C{alpha} had provided data on a small patient group, indicating a subset of patients exist with activated protein kinase C{alpha} (26). Although coincidence detection of EGFR activity can be employed in tumor cell lines using FRET efficiency, it is the automation of this method for significant patient numbers that is crucial for outcome analysis (27). Moreover, in these studies there were insufficient numbers to derive statistically significant predictions about disease progression. The developments described here provide a substantial advance that permit significant numbers of patient samples to be analyzed.

The two principal classes of EGFR inhibitors are monoclonal antibodies that bind to the extracellular domain (of which the most promising drug is the chimeric human:murine monoclonal antibody cetuximab) and quinazoline inhibitors of the intracellular kinase domain (of which the ones that are most commonly used are gefitinib and erlotinib; ref. 28). Gefitinib has been tested in a phase II trial in HNSCC and the results show a disease stabilization of 29% (29). Preliminary results from a phase III trial, which randomly assigned 424 head and neck cancer patients to receive either radiation therapy alone or radiation with concurrent cetuximab, have shown that cetuximab nearly doubled the median survival in these patients (30). However, the two phase II head and neck cancer trials on cetuximab revealed a relatively low response rate (around 10%) despite the fact that the majority of the patients have a 2 to 3+ staining for EGFR (31, 32). And in colon cancer two trials found no relation of EGFR expression and response rate to cetuximab (33, 34). In fact, one trial has reported objective responses to cetuximab in chemotherapy-refractory colorectal cancer even when EGFR expression was not detected by immunohistochemistry (35). Therefore, these trials on cetuximab indicate poor predictive value of EGFR expression in response to cetuximab. It would seem important in future studies to determine how EGFR activation status, rather than concentration, track with response to these and related EGFR-directed treatments.

Monitoring response to treatment is of increasing importance with respect to trials of new molecular-directed therapeutics, as well as in seeking to optimize the use of those reaching the clinic. It is well accepted that the use of biomarkers in the development of new agents is important and it is implicit that these markers reflect the action of the target (whether direct or through a validated surrogate) and not simply its presence. The logic applied to the targeting and monitoring of inhibition for these targets is no less relevant to the question of prognostic indicators in relation to disease. Indeed, the studies reported here clearly show the distinctive value of the information acquired in comparing the prognostically uninformative data set relating to two different means of assessing the expression pattern of EGFR levels in these HNSCC samples and the apparently informative data set obtained from determining the phosphorylation status of the receptor.

The need to be able to determine the functional status of potential therapeutic targets or prognostic indicators is irresistible. For the EGFR studied here, the ligand-induced autophosphorylation of tyrosine residues within the cytoplasmic domain serves as a useful marker of receptor activation that is stable with fixation and sample storage. Many such covalent modifications are well characterized functionally in this and other pathways. Hence, there is a significant opportunity to develop and exploit pairs of reagents to monitor these pathways in tissue microarrays using the type of highly selective, two-site, automated system described here.

For HNSCC, it is anticipated that FRET efficiency measurements by automated high-throughput FLIM will complement the traditional EGFR immunohistochemistry methods and the clinicopathologic information in aiding clinical decision and management of head and neck cancer patients. We recognize that our study is an exploratory assessment of disease-free survival. A prospective protocol-driven assessment for validation and determination of performance characteristics needs to be done using this potentially powerful technique as described by Hayes et al. (36, 37). Using this strategy, we would propose using this assay for prospective stratification of patients in randomized trials of EGFR inhibition; retrospective analysis of trials of EGFR inhibitors would also be possible. In the longer term, the identification of these patients at presentation should provide guidance on more aggressive treatment or, more specifically, on appropriate molecular therapy.


   Acknowledgments
 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. R. Byrne, J. Downward, and G. Stamp for critically reading the manuscript; Litia Campo for making the tumor microarrays; Dr. P. Chowdhury for her assistance; Dr. Gavin Kelly (Department of Bioinformatics and Biostatistics, Cancer Research UK, London Research Institute) for univariate and multivariate analysis of the prognostic factors using R.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

A. Kong and P. Leboucher contributed equally to this work.

Received 8/22/05. Revised 12/23/05. Accepted 12/30/05.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Hoffman HT, Karnell LH, Funk GF, Robinson RA, Menck HR. The National Cancer Data Base report on cancer of the head and neck. Arch Otolaryngol Head Neck Surg 1998;124:951–62.[Abstract/Free Full Text]
  2. Pisani P, Bray F, Parkin DM. Estimates of the world-wide prevalence of cancer for 25 sites in the adult population. Int J Cancer 2002;97:72–81.[CrossRef][Medline]
  3. Hardisson D. Molecular pathogenesis of head and neck squamous cell carcinoma. Eur Arch Otorhinolaryngol 2003;260:502–8.[CrossRef][Medline]
  4. Forastiere A, Koch W, Trotti A, Sidransky D. Head and neck cancer. N Engl J Med 2001;345:1890–900.[Free Full Text]
  5. Grandis JR, Tweardy DJ. Elevated levels of transforming growth factor {alpha} and epidermal growth factor receptor messenger RNA are early markers of carcinogenesis in head and neck cancer. Cancer Res 1993;53:3579–84.[Abstract/Free Full Text]
  6. Rubin Grandis J, Melhem MF, Gooding WE, et al. Levels of TGF-{alpha} and EGFR protein in head and neck squamous cell carcinoma and patient survival. J Natl Cancer Inst 1998;90:824–32.[Abstract/Free Full Text]
  7. Gupta AK, McKenna WG, Weber CN, et al. Local recurrence in head and neck cancer: relationship to radiation resistance and signal transduction. Clin Cancer Res 2002;8:885–92.[Abstract/Free Full Text]
  8. Ang KK, Berkey BA, Tu X, et al. Impact of epidermal growth factor receptor expression on survival and pattern of relapse in patients with advanced head and neck carcinoma. Cancer Res 2002;62:7350–6.[Abstract/Free Full Text]
  9. Dassonville O, Formento JL, Francoual M, et al. Expression of epidermal growth factor receptor and survival in upper aerodigestive tract cancer. J Clin Oncol 1993;11:1873–8.[Abstract/Free Full Text]
  10. Sheridan MT, O'Dwyer T, Seymour CB, Mothersill CE. Potential indicators of radiosensitivity in squamous cell carcinoma of the head and neck. Radiat Oncol Investig 1997;5:180–6.[CrossRef][Medline]
  11. Ulanovski D, Stern Y, Roizman P, Shpitzer T, Popovtzer A, Feinmesser R. Expression of EGFR and Cerb-B2 as prognostic factors in cancer of the tongue. Oral Oncol 2004;40:532–7.[Medline]
  12. Wen QH, Miwa T, Yoshizaki T, Nagayama I, Furukawa M, Nishijima H. Prognostic value of EGFR and TGF-{alpha} in early laryngeal cancer treated with radiotherapy. Laryngoscope 1996;106:884–8.[CrossRef][Medline]
  13. Parra HS, Cavina R, Latteri F, et al. Analysis of epidermal growth factor receptor expression as a predictive factor for response to gefitinib ("Iressa," ZD1839) in non-small-cell lung cancer. Br J Cancer 2004;91:208–12.[Medline]
  14. Dei Tos AP, EIlis I. Assessing epidermal growth factor receptor expression in tumours: What is the value of current test methods? Eur J Cancer 2005;41:1383–92.[CrossRef][Medline]
  15. Atkins D, Reiffen KA, Tegtmeier CL, Winther H, Bonato MS, Storkel S. Immunohistochemical detection of EGFR in paraffin-embedded tumor tissues: variation in staining intensity due to choice of fixative and storage time of tissue sections. J Histochem Cytochem 2004;52:893–901.[Abstract/Free Full Text]
  16. Chung CH, Parker JS, Karaca G, et al. Molecular classification of head and neck squamous cell carcinomas using patterns of gene expression. Cancer Cell 2004;5:489–500.[CrossRef][Medline]
  17. Bei R, Budillon A, Masuelli L, et al. Frequent overexpression of multiple ErbB receptors by head and neck squamous cell carcinoma contrasts with rare antibody immunity in patients. J Pathol 2004;204:317–25.[CrossRef][Medline]
  18. Arteaga CL. Epidermal growth factor receptor dependence in human tumors: more than just expression? Oncologist 2002;7 Suppl 4:31–9.[Abstract/Free Full Text]
  19. Larijani B, Allen-Baume V, Morgan CP, Li M, Cockcroft S. EGF regulation of PITP dynamics is blocked by inhibitors of phospholipase C and of the Ras-MAP kinase pathway. Curr Biol 2003;13:78–84.[CrossRef][Medline]
  20. Wu P, Brand L. Resonance energy transfer: methods and applications. Anal Biochem 1994;218:1–13.[CrossRef][Medline]
  21. Van de Vijver MJ, Kumar R, Mendelsohn J. Ligand-induced activation of A431 cell epidermal growth factor receptors occurs primarily by an autocrine pathway that acts upon receptors on the surface rather than intracellularly. J Biol Chem 1991;266:7503–8.[Abstract/Free Full Text]
  22. Brennan JA, Mao L, Hruban RH, et al. Molecular assessment of histopathological staging in squamous-cell carcinoma of the head and neck. N Engl J Med 1995;332:429–35.[Abstract/Free Full Text]
  23. Boyle JO, Hakim J, Koch W, et al. The incidence of p53 mutations increases with progression of head and neck cancer. Cancer Res 1993;53:4477–80.[Abstract/Free Full Text]
  24. Endo S, Zeng Q, Burke NA, et al. TGF-{alpha} antisense gene therapy inhibits head and neck squamous cell carcinoma growth in vivo. Gene Ther 2000;7:1906–14.[CrossRef][Medline]
  25. Todd R, Donoff BR, Gertz R, et al. TGF-{alpha} and EGF-receptor mRNAs in human oral cancers. Carcinogenesis 1989;10:1553–6.[Abstract/Free Full Text]
  26. Ng T, Squire A, Hansra G, et al. Imaging protein kinase C{alpha} activation in cells. Science 1999;283:2085–9.[Abstract/Free Full Text]
  27. Keese M, Magdeburg RJ, Herzog T, et al. Imaging epidermal growth factor receptor phosphorylation in human colorectal cancer cells and human tissues. J Biol Chem 2005;280:27826–31.[Abstract/Free Full Text]
  28. Mendelsohn J, Baselga J. Status of epidermal growth factor receptor antagonists in the biology and treatment of cancer. J Clin Oncol 2003;21:2787–99.[Abstract/Free Full Text]
  29. Cohen EE, Rosen F, Stadler WM, et al. Phase II trial of ZD1839 in recurrent or metastatic squamous cell carcinoma of the head and neck. J Clin Oncol 2003;21:1980–7.[Abstract/Free Full Text]
  30. Bonner JA, Giralt J, Harari PM, et al. Cetuximab prolongs survival in patients with locoregionally advanced squamous cell carcinoma of head and neck: a phase III study of high dose radiation therapy with or without cetuximab [abstract]. Proc Am Soc Clin Oncol 2004;22:5507.
  31. Baselga J, Trigo JM, Bourhis J, et al. Phase II multicenter study of the antiepidermal growth factor receptor monoclonal antibody cetuximab in combination with platinum-based chemotherapy in patients with platinum-refractory metastatic and/or recurrent squamous cell carcinoma of the head and neck. J Clin Oncol 2005;23:5568–77.[Abstract/Free Full Text]
  32. Herbst RS, Arquette M, Shin DM, et al. Phase II multicenter study of the epidermal growth factor receptor antibody cetuximab and cisplatin for recurrent and refractory squamous cell carcinoma of the head and neck. J Clin Oncol 2005;23:5578–87.[Abstract/Free Full Text]
  33. Saltz LB, Meropol NJ, Loehrer PJ, Sr., Needle MN, Kopit J, Mayer RJ. Phase II trial of cetuximab in patients with refractory colorectal cancer that expresses the epidermal growth factor receptor. J Clin Oncol 2004;22:1201–8.[Abstract/Free Full Text]
  34. Cunningham D, Humblet Y, Siena S, et al. Cetuximab monotherapy and cetuximab plus irinotecan in irinotecan-refractory metastatic colorectal cancer. N Engl J Med 2004;351:337–45.[Abstract/Free Full Text]
  35. Chung KY, Shia J, Kemeny NE, et al. Cetuximab shows activity in colorectal cancer patients with tumors that do not express the epidermal growth factor receptor by immunohistochemistry. J Clin Oncol 2005;23:1803–10.[Abstract/Free Full Text]
  36. Hayes DF, Bast RC, Desch CE, et al. Tumor marker utility grading system: a framework to evaluate clinical utility of tumor markers. J Natl Cancer Inst 1996;88:1456–66.[Abstract/Free Full Text]
  37. Hayes DF, Trock B, Harris AL. Assessing the clinical impact of prognostic factors: when is "statistically significant" clinically useful? Breast Cancer Res Treat 1998;52:305–19.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 JCOHome page
Y. Nieto, F. Nawaz, R. B. Jones, E. J. Shpall, and S. Nawaz
Prognostic Significance of Overexpression and Phosphorylation of Epidermal Growth Factor Receptor (EGFR) and the Presence of Truncated EGFRvIII in Locoregionally Advanced Breast Cancer
J. Clin. Oncol., October 1, 2007; 25(28): 4405 - 4413.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
A. A. Molinolo, S. M. Hewitt, P. Amornphimoltham, S. Keelawat, S. Rangdaeng, A. Meneses Garcia, A. R. Raimondi, R. Jufe, M. Itoiz, Y. Gao, et al.
Dissecting the Akt/Mammalian Target of Rapamycin Signaling Network: Emerging Results from the Head and Neck Cancer Tissue Array Initiative
Clin. Cancer Res., September 1, 2007; 13(17): 4964 - 4973.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kong, A.
Right arrow Articles by Larijani, B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kong, A.
Right arrow Articles by Larijani, B.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online