Inhibition of epidermal growth factor receptor gene ... - Nature

Oncogene (1997) 15, 409 ± 416  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Inhibition of epidermal growth factor receptor gene expression and function decreases proliferation of head and neck squamous carcinoma but not normal mucosal epithelial cells J Rubin Grandis1, A Chakraborty3, MF Melhem2, Q Zeng1 and DJ Tweardy3 Departments of 1Otolaryngology, 2Pathology, 3Medicine and Molecular Genetics and Biochemistry, University of Pittsburgh and the University of Pittsburgh Cancer Institute, Pittsburgh 15213, USA

Previous reports have shown that fresh tissues and cell lines from patients with squamous cell carcinoma of the head and neck (SCCHN) overexpress transforming growth factor alpha (TGF-a) and its receptor, the epidermal growth factor receptor (EGFR) at both the mRNA and protein levels. Protein localization studies con®rm that TGF-a and EGFR are produced by the same epithelial cells in tissues from head and neck cancer patients further supporting an autocrine growth pathway. Using three strategies, we examined the hypothesis that downmodulation of EGFR would reduce the proliferation of SCCHN cells. We targeted EGFR mRNA using antisense oligonucleotides and the mature EGFR protein at two sites, the ligand-binding domain and the kinase domain, and determined the eects of this targeting on SCCHN proliferation. Treatment of several SCCHN cell lines with a pair of antisense oligodeoxynucleotides directed against the translation start site and ®rst intronexon splice junction of the human EGFR gene resulted in decreased EGFR protein production and inhibited growth by 86% compared to a 13% reduction in cells treated with sense oligonucleotides (P=0.03). Growth inhibition was speci®c for carcinoma cells since the same EGFR antisense oligonucleotides had no eect on the proliferation of normal mucosa cells harvested from non-cancer patients. Two monoclonal antibodies which block ligand binding to EGFR (MAbs 425 and 528) inhibited the growth of several SCCHN cell lines by up to 97% which suggests that EGFR is participating in an autocrine pathway in SCCHN that is, at least in part, external. An EGFR-speci®c tyrosine kinase inhibitor (PD 153035) was found to inhibit EGFR phosphorylation in SCCHN cell lines and to reduce growth by 68% although it had no eect on the growth rate of normal mucosal epithelial cells. These experiments indicate that EGFR gene expression and function is critical for SCCHN cell growth but not for growth of normal mucosa cells and therefore may serve as a tumor-speci®c target for preventive and therapeutic strategies in head and neck cancer. Keywords: epidermal growth factor receptor; head and neck cancer

Correspondence: JR Grandis, The Eye & Ear Institute Building, Suite 500, 200 Lothrop Street, Pittsburgh, PA 15213 Received 17 January 1997; revised 2 April 1997; accepted 3 April 1997

Introduction EGFR belongs to the tyrosine kinase family of cell surface receptors. Members of the EGFR family are frequently implicated in human carcinomas. Overexpression of the wild-type EGFR gene under appropriate conditions in mammalian cells confers the transformed phenotype (DiMarco et al., 1989; DiFore and Krause, 1992). EGFR is upregulated in numerous epithelial malignancies, the most well-studied being breast cancer where its expression is correlated with both a poor prognosis and a lack of response to endocrine therapy (Sainsbury et al., 1987; Nicholson et al., 1988) as well as increased metastatic potential (Toi et al., 1991). In patients with colon carcinoma, increased EGFR mRNA in the tumors is associated with a higher rate of liver metastasis (Radinsky et al., 1995). Elevated EGFR has also been detected in cancers of the bladder (Harris and Nicholson, 1989), lung (Veale et al., 1987), kidney (Atlas et al., 1992) and ovary (Morishige et al., 1991). We and others have shown that EGFR mRNA and protein is overexpressed in tumors and cell lines obtained from patients with SCCHN (Ozanne et al., 1986; Maxwell et al., 1989; Kawamoto et al., 1991; Christensen et al., 1992; Rubin Grandis and Tweardy, 1993). While EGFR gene ampli®cation has been reported in selected SCCHN cell lines and tumor samples (Yamamoto et al., 1986; Weichselbaum et al., 1989; Ishitoya et al., 1989), most of the evidence suggests that upregulation detected at the mRNA and protein level cannot be accounted for by an increase in gene dosage alone (Eisbruch et al., 1987; Rikimaru et al., 1992). We recently reported that activation of EGFR gene transcription appears to be the primary mechanism of overexpression in SCCHN as opposed to gene dosage and mRNA halflife) (Rubin Grandis et al., 1996a). Elevated EGFR has been associated in several cases with larger tumor size, advanced stage and hence, a worse prognosis (Todd et al., 1989). Downmodulation of EGFR expression in human carcinoma cell lines using antisense constructs has resulted in suppression of the transformed phenotype and restoration of serum and anchorage-dependent growth (Moroni et al., 1992). Others have reported decreased tumor cell proliferation in vitro and tumor growth in vivo of EGFR over-producing cells treated with EGFR antibodies which block ligand-binding (Fan et al., 1993; Baselga and Mendelsohn, 1994a, b). EGFR-directed therapy has been used in patients with malignancies known to overexpress EGFR protein. Such treatment has included monoclonal

EGFR inhibition reduces head and neck cancer cell proliferation JR Grandis et al

410

antibodies against EGFR for glioblastomas (Derui et al., 1992) and carcinomas of the breast (Baselga et al., 1993) or lung (Divgi et al., 1991) alone or in combination with chemotherapy. Also, toxins elaborated by Pseudomonas or Diphtheria species have been chemically linked to TGF-a and/or EGF allowing them to be targeted to cells overexpressing EGFR for therapeutic application (LeMaistre et al., 1992; Theuer et al., 1993; Phillips et al., 1994; Kameyama et al., 1994). The present investigation was designed to test the hypothesis that EGFR gene expression is critical for SCCHN proliferation by examining the eects of downmodulating EGFR on cell growth. Using three strategies to decrease EGFR expression and function including antisense oligonucleotides, monoclonal antibodies directed against the ligand-binding domain of EGFR (MAbs 425 and 528), and an EGFR-speci®c tyrosine kinase inhibitor (PD153035), we demonstrated that this downmodulation was accompanied by decreased proliferation of SCCHN cells, while control normal mucosal epithelial cells remained unaected.

Results EGFR protein is produced by SCCHN but not control cells We have previously shown that EGFR protein is produced by squamous epithelial cells in tissue sections of both histologically normal mucosa and tumor tissue from patients with SCCHN. In contrast, EGFR protein was nearly undetectable in mucosal epithelium from control patients without cancer (Rubin Grandis et al., 1996b). In this study we examined nine SCCHN cell lines known to over-produce EGFR mRNA as well as normal mucosal epithelial cells from non-cancer patients for EGFR protein expression using an EGFR ELISA assay. All nine SCCHN cell lines examined expressed EGFR protein while there was minimal EGFR protein detected in the normal mucosa cells (Figure 1). EGFR antisense oligonucleotide treatment decreased EGFR protein expression and cell proliferation in SCCHN cells but had no eect on control normal mucosa cells from non-cancer patients To determine whether inhibition of expression of the EGFR gene would reduce SCCHN cell proliferation, several SCCHN cell lines were treated with a pair of EGFR-speci®c antisense oligonucleotides directed against the translation start site and ®rst splice-donor acceptor site of human EGFR and their growth rates were determined. SCCHN cells (PCI 37a) grown in the presence of sense oligonucleotides grew at a rate which was comparable to the number of cells on day 10 which had not received oligonucleotides (Figure 2a). In contrast, cells grown in the presence of EGFR-speci®c antisense oligonucleotides revealed a mean 80% growth inhibition on day 6 compared with a 13% growth inhibition in sense oligonucleotide-treated cells (P=0.03). Similar results were obtained with several other SCCHN cell lines (PCI 15b, Figure 2b and data

Figure 1 EGFR ELISA assay of nine SCCHN cell line and control mucosal epithelial cell lysates. The data are the mean+s.d. of two experiments. On average, 1 mg protein was obtained from 66106 cells. Consequently, 1000 fmol/mg represents 105 EGFR per cell

not shown; PCI-13, UPCI-SCC-66, UPCI-SCC-104, 1483). In contrast, normal mucosa cells harvested from non-cancer patients were not growth-inhibited by EGFR antisense oligonucleotide treatment (Figure 2c) suggesting that the EGFR antisense eects are speci®c for carcinoma cells. To determine that the inhibition of SCCHN cell proliferation by antisense oligonucleotides resulted from a decrease in EGFR on the cell surface, FACS analysis was performed of SCCHN cells treated with oligonucleotides using MAb 425 as the primary antibody. Flow cytometry demonstrated a 68% reduction in mean EGFR staining in cells treated with antisense oligonucleotides compared with cells which received no oligonucleotides (Figure 3a). To con®rm that EGFR antisense oligonucleotide treatment reduced the amount of EGFR protein, we repeated the EGFR ELISA following 4 days of treatment with antisense oligonucleotides and found an 87% reduction in EGFR protein levels in SCCHN cells, (PCI 15b) which received antisense therapy (Figure 3b). Monoclonal antibodies which block EGFR ligand-binding inhibited the growth of SCCHN but not normal mucosal epithelial cells To dierentiate between an internal and external autocrine pathway we exposed several SCCHN cell lines and control cells to increasing concentrations (0 ± 10 nM) of two dierent EGFR-speci®c monoclonal antibodies (MAbs 425 and 528) known to block ligand-binding (Figure 4 and data not shown; PCI-13, UPCI-SCC-66, UPCI-SCC-104, 1483). Both antibodies resulted in an inhibition of cell proliferation of SCCHN cell lines (96.3 and 97%, respectively; PCI 15b and PCI 37a), but not normal mucosal epithelial cell proliferation.


411

a b

c

Figure 2 Eect of antisense oligonucleotides against EGFR on SCCHN and normal epithelial cell growth. (a) a representative SCCHN cell line (PCI 37a) was growth inhibited by incubation with EGFR antisense oligonucleotides. Cells were exposed to antisense (^), sense (&), or no oligonucleotides (*) and were harvested and counted at the times indicated. (b) shows the same experiment using a dierent SCCHN cell line (PCI 15b). (c) the growth rate of normal mucosal epithelial cells treated with EGFR antisense oligonucleotides (^) is similar to sense-treated (&) and control (*) cells. Cell counts were performed up to ten days or when the cells became con¯uent

EGFR tyrosine kinase inhibitor treatment of SCCHN but not control normal mucosa cells results in decreased growth rate via inhibition of receptor phosphorylation The eect of an EGFR-speci®c tyrosine kinase inhibitor (PD 153035) on several SCCHN cell lines was examined by [3H]thymidine incorporation assays and con®rmed by cell counting studies at various time points after the addition of the inhibitor (Figure 5 and data not shown; PCI-13, UPCI-SCC-66, UPCI-SCC104, 1483). The incorporation of [3H]thymidine was reduced by a mean of 68% following treatment with 300 nM of the inhibitor (Figure 5a). In the time-course studies, cell number was reduced by a mean of 45% at day 6 in the SCCHN cells plated in PD 153035containing medium compared with controls cells plated in medium alone (Figure 5b). The growth rate of

normal mucosal epithelial cells was unaected by the EGFR-speci®c tyrosine kinase inhibitor (Figure 5c). To con®rm that the mechanism of growth-suppression in SCCHN cells by inhibitor involved inhibition of EGFR tyrosine kinase activity, we investigated the level of tyrosine phosphorylation of EGFR and EGFRassociated proteins by performing immunoblotting of immunoprecipitated EGFR samples with anti-phosphotyrosine antibody (Figure 6). Twenty mM of inhibitor signi®cantly reduced ligand-induced EGFR phosphorylation and 100 mM abolished it. Discussion These studies demonstrate that downmodulation of the increased EGF receptors produced by head and neck


412

a

a

b b

Figure 3 (a) ¯ow cytometry analysis of SCCHN cells (PCI 15b) treated with EGFR sense (.....) or antisense (- - - -) oligonucleotides or receiving no oligonucleotide treatment (ÐÐ) followed by staining with MAb 425. (b) EGFR ELISA assay demonstrated an 87% decrease in EGFR protein in SCCHN cells ) (PCI 15b) treated with EGFR antisense oligonucleotides ( ) or no compared with cells which received sense ( ) oligonucleotides (

cancer cells using three approaches including antisense oligonucleotides or monoclonal antibodies, and an EGFR-speci®c tyrosine kinase inhibitor, all resulted in decreased SCCHN cell proliferation. These modulating agents targeted EGFR at three distinct biochemical sites including mRNA and protein synthesis, ligand binding to receptor and receptor tyrosine kinase activity. The eect of each appeared to be speci®c for carcinoma cells since non-transformed mucosal epithelial cells were not inhibited by any of these three reagents. These results suggest that EGFR gene expression is critical for proliferation in SCCHN cells but not in normal mucosal epithelial cells. Upregulation of EGFR and its ligand, transforming growth factor alpha (TGF-a), appears to be an early

Figure 4 Eect of EGFR-speci®c monoclonal antibodies on SCCHN cell growth. (a) represents the eects of increasing concentrations of MAb 425 on the proliferation of SCCHN cell ), PCI-15b ( ) and normal mucosal epithelial lines PCI-37a ( ). (b) represents the eects of increasing concentrations cells ( of MAb 528 on the same cells. The data are the mean+s.d. of three experiments

event in head and neck carcinogenesis since elevated EGFR levels are detected in the histologically normal mucosa from SCCHN patients in contrast to nearly undetectable levels in control normal mucosa from non-cancer patients (Rubin Grandis and Tweardy, 1993; Rubin Grandis et al., 1996b). We and others have also found increasing EGFR protein in mucosa that is in closer proximity to the tumor as well as higher expression associated with increasing degrees of epithelial dysplasia (Rubin Grandis et al., 1996b; Shin et al., 1994). These cumulative ®ndings support a correlation between activation of EGFR gene transcription and head and neck oncogenesis. The elevated EGFR expression may also be clinically signi®cant since we have recently noted decreased survival in


413

a

b

c

Figure 5 EGFR-speci®c tyrosine kinase inhibitor (PD 153035) eects on SCCHN cell line (PCI-37a) proliferation. (a) represents the results of duplicate [3H]thymidine incorporation assays. Cells were rested in SFM for 24 h followed by addition of PD 153035 for 72 h with an 18 h pulse with [3H]thymidine. (b) represents duplicate cell count experiments comparing the eects of the inhibitor (&) on SCCHN cells or no treatment (^). (c) shows the results of duplicate cell count experiments comparing the EGFR inhibitor (&) to diluent alone (control) (^) on the growth rate of control mucosal epithelial cells

SCCHN patients whose primary tumors express dramatically increased EGFR protein levels (Rubin Grandis et al., unpublished data). We designed a pair of antisense oligonucleotides against the translation start site and ®rst splice donoracceptor site of the human EGFR gene and found that several SCCHN cell lines were growth inhibited by treatment with these antisense oligonucleotides. Subsequent analysis revealed that this growth inhibition was due, at least in part, to decreased EGFR protein expression on the cell surface. The eect of EGFR antisense oligonucleotides was also examined on normal mucosa cells harvested from non-cancer patients with no modulation of growth resulting from antisense oligonucleotide treatment. This suggests that the upregulated EGFR drives proliferation in SCCHN

cells but not in the corresponding normal mucosa epithelial cell. We observed similar results when the eects of TGF-a-speci®c antisense oligonucleotides were examined in SCCHN and corresponding normal cells (Rubin Grandis et al., unpublished observations) providing further evidence of an autocrine pathway involving TGF-a and EGFR in SCCHN but not in corresponding normal upper aerodigestive tract epithelium that is not at high risk for transformation. In an attempt to modulate EGFR by blocking receptor kinase activity, we used an EGFR-speci®c tyrosine kinase inhibitor (PD 153035) that has not previously been shown to inhibit the growth of upper aerodigestive tract squamous carcinoma cells that upregulate EGFR (Fry et al., 1994), and found that the inhibitor decreased proliferation of SCCHN but


414

not of corresponding normal mucosal epithelial cells. Further studies showed that inhibition of SCCHN cell growth appeared to result from inhibition of EGF receptor phosphorylation. These ®nidngs suggests that EGFR-speci®c tyrosine kinase activity is necessary for SCCHN cell growth but not the growth of control normal mucosa cells. Moreover, the addition of exogenous ligand (EGF) was able to increase EGFR phosphorylation which indicates that despite upregulation of EGFR, unbound receptors remain available on the surface of SCCHN cells to bind exogenous ligand. Previous studies in our laboratory have demonstrated that TGF-a which is produced by SCCHN cell lines remains primarily cell-associated (Rubin Grandis, unpublished observations). The ligand-receptor interaction, therefore, may occur either in the cytoplasm, as has been previously suggested (Massague, 1990;

EGF Inhibitor

– –

+ –

+ 20

+ 100

203 — — EGFR 118 —

203 — — EGFR

Laird et al., 1994; Kenney et al., 1993), or on the cell surface. To dierentiate between an intracrine and a juxtacrine or external autocrine pathway, we treated SCCHN cells with two dierent monoclonal antibodies known to block EGFR ligand-binding. Our results suggest that TGF-a and EGFR are participating, at least in part, in an external autocrine growth pathway in most of the SCCHN cell lines since treatment with either antibody resulted in partial SCCHN growth-inhibition. Failure to completely inhibit SCCHN cell growth with EGFR antibody treatment raises the possibility that the autocrine pathway involving TGF-a and EGFR may be partially internal. If there is both an internal as well as an external autocrine pathway, then treatment strategies which target this loop may require intracellular delivery to optimize ecacy. Receptor and ligand-directed therapy has been explored in several malignancies which overexpress EGFR protein. Such treatment has included monoclonal antibodies against EGFR in lung and breast cancer (Baselga et al., 1993; Divgi et al., 1991). Fusion proteins that link TGF-a to toxins by Pseudomonas or Diphtheria species (Phillips et al., 1994; Kameyama et al., 1994) have been shown to be eective in inhibiting tumor growth in xenograft models (Kihara and Pastan, 1994). Recent data suggests that EGFR tyrosine kinase inhibitors may be ideal candidates for cancer chemopreventive studies (Kello et al., 1996). Our current results demonstrating that downmodulation of EGFR speci®cally results in inhibition of SCCHN cell growth combined with our preliminary ®nding of a survival advantage for patients whose SCCHN tumors express lower levels of EGFR protein suggests that receptor-directed treatment strategies merit further evaluation in SCCHN.

118 —

Materials and methods Figure 6 EGFR-speci®c tyrosine kinase inhibitor eects on EGFR tyrosine phosphorylation in SCCHN cells. In the upper panel, SCCHN cells (PCI-37a) were incubated in media without or with EGF (100 ng/ml) and without or with the tyrosine kinase inhibitor PD153035 at the concentration (mM) indicated. Cells were lysed and the EGFR immunoprecipitated from cell lysates. Following separation on SDS ± PAGE and transfer to membrane, the blots were probed with antiphosphotyrosine antibody. In the lower panel, a duplicate blot was probed with an anti-EGFR monoclonal antibody (MAb425)

Cells Cell lines derived from patients with SCCHN were grown in Dulbecco's Modi®ed Eagle's Medium (DMEM) (Cellgro, Washington, DC) with 15% fetal bovine serum (GIBCO Laboratories, Grand Island, NY), plus 100 units/ ml of penicillin and 100 units/ml of streptomycin (GIBCO) (Saachi et al., 1990). Most of the SCCHN cell lines are part of a large collection of established cell lines in the

Table 1 Characteristics of SCCHN Cell Lines Cell Line

Primary Site

Gender/Age

Cultured Organ Site

TNM Stage

PCI-1

Larynx

M/65

N/A

PCI-13

Retromolar trigone Cervical lymph node Larynx Larynx Alveolar ridge Floor of mouth Cervical lymph node Retromolar trigone

M/50

Recurrent laryngeal tumor Primary tumor Metastatic lymph node Primary tumor Primary tumor Primary tumor Primary tumor Metastatic lymph node Primary tumor

T3N1M0

PCI-15B PCI-37A PCI-52 UPCI-SCC-66 UPCI-SCC-104 UM-SCC-22B 1483

M/69 F/62 M/43 F/75 M/57 F/59 M/66

T3N1M0

T3N2M0 T2N0M0 T1N0M0 TyNxM0 T2N1M0 T2N1M0


Department of Otolaryngology at the University of Pittsburgh (Heo et al., 1989). The SCCHN cell line 22B was kindly provided by Dr Thomas Carey (University of Michigan) and 1483 was generously provided by Dr Reuben Lotan (MD Anderson Cancer Center). The SCCHN cell lines UPCI:SCO66 and UPCI:SCC104 were a generous gift from Dr Susanne M Gollin (University of Pittsburgh) (Table 1). Established cell lines are obtained in early passage, expanded, and frozen. Cells for each experiment were thawed and used as described. Primary mucosal cultures were established from oropharyngeal mucosa harvested from a series of control patients without cancer as described previously (Rubin Grandis et al., 1996a). EGFR protein detection Cell lysates were prepared from untreated SCCHN cell lines and SCCHN cell lines treated with oligonucleotides as well as primary mucosal cultures from non-cancer patients according to the manufacturer's instructions. Total protein was quantitated using the BioRad Assay method and 50 mg of total protein per sample was analysed using an EGFR ELISA (lower limits of detection of the assay is 5 fmol/ml of EGFR or EGF receptor extracellular domain according to the manufacturer, Oncogene Science). Antisense oligodeoxynucleotide studies Phosphorothioated 18 and 23-mer oligodeoxynucleotides were synthesized on an Applied Biosystem 394 DNA synthesizer by means of B-cyanothylphosphoramidite chemistry to minimize degradation by endogenous nucleases. The ®rst antisense oligonucleotide (5'-CGGA GGGTCGCATCGCTG-3') was directed against the translation start site (AUG codon) and surrounding nucleotides of the human EGFR gene. The sense oligonucleotide was 5'-CAGCGATGCGACCCTCCG-3' (1109 ± 1126). The second antisense oligonucleotide targeted the ®rst intron-exon splice junction (5'-CACGCCCTTACCTTTCTTTTCCT-3'). The corresponding sense oligonucleotide was 5'-AGGAAAAGAAAGGTAAG GGCGTG-3' (1190 ± 1212) (Haley et al., 1987). To examine the eect of oligonucleotide treatment on proliferation, SCCHN or control mucosal cells previously shown to take up approximately 40% of oligonucleotide by 24 h (Rubin Grandis et al., unpublished observations), were plated at a density of 104 cells/microtiter well in 24-well polystyrene plates (Falcon). After 24 h, at which point the cells had reached 30 ± 40% con¯uency, 250 ml of fresh culture medium containing 12.5 mM antisense oligonucleotide targeted against the translation start site and 12.5 mM antisense oligonucleotide directed against the splice-donor acceptor site of human EGFR was added. As control, cells received an equal concentration of corresponding EGFR sense oligonucleotides or the same medium without oligonucleotides. An additional 100 ml of fresh culture medium containing 12.5 mM of each of the two antisense or sense oligonucleotides or medium containing no oligonucleotides was added to the wells on days 3, 6 and 9. Cell counts were performed over a period of 10 days, with duplicate samples for each time point. The percentage of growth inhibition was calculated by dividing the dierence in cell number between the cells treated with sense oligonucleotides and those treated with antisense oligonucleotides by the number of cells in the senseoligonucleotide-treated group. To determine the eect of EGFR antisense oligonucleotides on EGFR protein production, SCCHN cells were assayed using the EGFR ELISA assay as described above following 4 days of oligonucleotide treatment. To con®rm the reduction in EGFR following oligonucleotide exposure,

SCCHN cells were plated in 96 well round bottom microtiter plates at a concentration of 2.56105 cell/well and labeled with primary antibody (MAb 425; 5 mg/ml) for 1 h at 48C then washed several times with phosphate buered saline containing 1% BSA then labeled with secondary antibody (FITC-conjugated goat anti-mouse; 1 mg/ml) for 1 h. Cells were then washed and analysed immediately on a ¯ow cytometer (Becton Dickinson FACStar). Proliferation assays SCCHN cell lines were plated in quadruplicate in 96-well ¯at bottomed polystyrene plates at a density of 16104 cells/well in serum free media for 24 h. Dierent concentrations (0.01 nM to 1 mM) of an EGFR-speci®c tyrosine kinase inhibitor (PD-153035, kindly provided by Dr David W Fry, Parke-Davis Pharmaceutical Research) in 1% serum containing media (DMEM with 1% pen/ strep) were added. Cells were harvested after 72 h following an 18 h pulse with [3H]thymidine and counts determined on a scintillation counter. The response of each cell line to dierent concentrations of serum (0 ± 10%) was measured each time as a positive control for proliferation. The eect of the EGFR-speci®c tyrosine kinase inhibitor on cell growth was also assessed by plating SCCHN cells in duplicate in 24-well polystyrene plates (Falcon) and performing cell counts by Erythrosin B dye exclusion 2, 4, 6 and 8 days following addition of the inhibitor. EGFR antibody studies A murine monoclonal antibody to EGFR (MAb 425) was generously provided by Dr Uli Rodeck. MAb 425 is of IgG2a isotype and binds to a protein determinant on the external domain of the human EGF receptor, inhibits binding of EGF, induces down-regulation of the EGF receptor, and does not stimulate EGFR kinase activity (Murthy et al., 1990). In vitro growth inhibition studies were carried out as previously described (Rodeck et al., 1987). All studies were repeated using a second EGFR MAb (528), also known to block ligand binding (Baselga and Mendelsohn, 1994a, b) (Oncogene Science). SCCHN and normal mucosal epithelial cells were plated in 24-well plates at a density of 36104 cells/well in serum-freemedium. After 24 h, fresh medium supplemented with 1% fetal calf serum and increasing concentrations of MAb 425 or MAb 528 were added in duplicate (0 ± 10 nM). After 5 days, the cells were trypsinized, resuspended and counted using a hemocytometer. Viability was assessed by Erythrosin B dye exclusion. In vitro kinase experiments A representative SCCHN cell line (PCI-37a) was grown to 75 ± 90% con¯uency, serum starved for 12 h and then incubated with increasing doses of the EGFR-speci®c tyrosine kinase inhibitor (0 ± 100 nM) for 2 h. Cells were then stimulated with recombinant EGF (100 ng/ml) for 5 min at 378C and then harvested and lysed in lysis buer (50 mM HEPES, pH 7.5; 150 mM NaCl; 1% TritonX-100, 1 m M Na 3 VO 4 , 10 m M sodium pyrophosphate; 50 m M NaF; 1 mM EDTA; 1 mM EGFT; 1 mM PMSF; 10 mg/ml of aprotinin and leupeptin) at 48C. After removal of cell debris by centrifugation, lysates were incubated with a murine monoclonal antibody against EGFR (MAb 425; 5 mg) for 4 h on ice. Protein A-Sepharose (40 ml of a 50% slurry; Pharmacia) was then added followed by 1 h of incubation with occasional gentle agitation. The immune complexes were washed thoroughly three times with lysis buer and resuspended in Laemli's sample buer and then

415


416

heated for 5 min at 1008C and electrophoresed in 10% SDS ± PAGE. Separated proteins were transferred in PVDF membrane (Milipore) blotted with an antiphosphotyrosine (4G10, Upstate Biotechnology) or EGFR antibody and bands were visualized with a chemiluminiscence kit (Amersham). Statistical analysis Statistical analysis of data was performed using Student's t-test. A probability of less than 5% (P50.05) was considered signi®cant.

Acknowledgements We are grateful to Dr Susanne M Gollin for providing us with head and neck cancer cell lines UPCI:SCC066 and UPCI:SCC104, to Dr Thomas Carey for providing the cell line 22B and Dr Reuben Lotan for the gift of the cell line 1483. We are also grateful to Dr David Fry for providing us with the EGFR kinase inhibitor PD-153035 and Dr Uli Rodeck for the gift of MAb 425. Supported in part by grants 1 K11 CA01760-01 from the National Cancer Institute, American Cancer Society grant DHP-111, the John R McCune Charitable Trust Foundation and the Mary Hillman Jennings Foundation.

References Atlas I, Mendelsohn J, Baselga J, Fair WR, Masui H and Kumar R. (1992). Cancer Res., 2, 3353 ± 3339. Baselga J and Mendelsohn J. (1994a). Pharmac. Ther., 64, 127 ± 154. Baselga J and Mendelsohn J. (1994b). Breast Cancer Res. and Treatment, 29, 127 ± 138. Baselga J, Norton L, Masui H, Pandiella A, Coplan K, Miller WH and Mendelsohn J. (1993). J. Natl. Cancer Inst., 85, 1327 ± 1333. Derui L, Woo V, Emrich J, Steplewski Z, Rodeck U, Herlyn D, Koprowski H, Miyamoto C and Brady LW. (1992). Am. J. Clin. Oncol., 15, 288 ± 294. Divgi CR, Welt S, Kris M, Real FX, Yeh SD, Gralla R, Merchant B, Schweighart S, Unger M, Larson SM et al. (1991). J. Natl. Cancer Inst., 83, 97 ± 104. Christensen ME, Therkildsen MH, Hansen BL, Albeck H, Hansen GN and Bretlau P. (1992). Eur. Arch. Otorhinolaryngol, 249, 243 ± 247. DiFore PP and Krause MH. (1992). Cancer Treat. Res., 61, 139 ± 160. DiMarco E, Pierce JH, Fleming TP, Kraus MH, Molloy CH, Aaronson SA and DiFiore PP. (1989). Oncogene, 4, 831 ± 838. Eisbruch A, Blick M, Lee JS, Sacks PG and Gutterman J. (1987). Cancer Res., 47, 3603 ± 3605. Fan Z, Masui H, Atlas I and Mendelsohn J. (1993). Cancer Res., 53, 4322 ± 4328. Fry DW, Kraker AJ, McMichael A, Ambroso LA, Nelson JM, Leopold WR, Connors RW and Bridges AJ. (1994). Science, 265, 1093 ± 1095. Haley J, Whittle N, Bennett P, Kinchington D, Ullrich A and Water®eld M. (1987). Oncogene Res, 1, 375 ± 396. Harris AL and Nicholson S. (1989). Cancer Treat. Res., 7, 353 ± 357. Heo DS, Snyderman CH, Gollin SM, Walker PE, Deka R, Barnes EL, Johnson JT, Herberman RB and Whiteside TL. (1989). Cancer Res., 49, 5167 ± 5175. Ishitoya J, Toriyama O, Oguchi N, Kitamura K, Ohshima M, Asano K and Yamamoto T. (1989). Br. J. Cancer, 59, 559 ± 562. Kameyama S, Kawamata H, Pastan I and Oyasu R. (1994). J. Cancer Res. Clin. Oncology, 120, 507 ± 512. Kawamoto T, Takahashi K, Nishi M, Kimura T, Matsumura T and Taniguchi S. (1991). Jpn. J. Cancer Res., 82, 403 ± 410. Kello GJ, Fay JR, Steele VE, Lubet RA, Boone CW, Crowell JA and Sigman CC. (1996). Cancer Epidem. Biomarkers & Prevention, 5, 657 ± 666. Kenney NJ, Saeki T, Gottardis M, Kim N, GarciaMorales P, Martin MB, Normanno N, Ciardiello F, Day A, Cutler ML et al. (1993). J. Cell. Physiol., 156, 497 ± 514. Kihara A and Pastan I. (1994). Bioconjugate Chemistry, 532 ± 538. Laird AD, Brown PL and Fausto N. (1994). Cancer Res., 54, 4224 ± 4232.

LeMaistre CF, Meneghetti C, Rosenblum M, Reuben J, Parker K, Shaw J, Deisseroth A, Woodworth T and Parkinson DR. (1992). Blood, 82, 444a. Massague J. (1990). J. Biol. Chem., 265, 21393 ± 21396. Maxwell SA, Sack PG, Gutterman JU and Gallick GE. (1989). Cancer Res., 49, 1130 ± 1137. Morishige K, Kurachi H, Amemiya K, Fukita Y, Yamamoto T, Miyake A and Tanizawa O. (1991). Cancer Res., 51, 5322 ± 5328. Moroni MC, Willingham MC and Beguinot L. (1992). J. Biol. Chem., 267, 2714 ± 2722. Murthy U, Rieman DJ and Rodeck U. (1990). Biochem. Biophy. Res. Comm., 172, 471 ± 476. Nicholson S, Halcrow P, Sainsbury JRC, Angus B, Chambers P, Farndon JR and Harris AL. (1988). Br. J. Cancer, 58, 810 ± 814. Ozanne B, Richards CS, Hendler F, Burns D and Gusterson B. (1986). J. Pathol., 149, 9 ± 14. Phillips PC, Levov C, Catterrall M, Colvin OM, Pastan I and Brem H. (1994). Cancer Res., 54, 1008 ± 1015. Radinsky R, Risin S, Fan D, Dong Z, Bielenberg D, Bucana CD and Fidler IJ. (1995). Clin. Cancer Res., 1, 19 ± 31. Rikimaru K, Tadokoro K, Yamamoto T, Enomoto S and Tsuchida N. (1992). Head & Neck, 14, 8 ± 13. Rodeck U, Herlyn M, Herlyn D, Moltho C, Atkinson B, Varello M, Steplewski L and Koprowski H. (1987). Cancer Res., 47, 3692 ± 3696. Rubin Grandis J and Tweardy DJ. (1993). Cancer Res., 53, 3579 ± 3584. Rubin Grandis J, Zeng Q and Tweardy DJ. (1996a). Nature Med., 2, 237 ± 240. Rubin Grandis J, Melhem MF, Barnes EL and Tweardy DJ. (1996b). Cancer, 78, 1284 ± 1292. Sacchi M, Snyderman CH, Heo DS, Johnson JT, d'Amico F, Herberman RB and Whiteside TL. (1990). Cancer Res., 50, 3113 ± 3118. Sainsbury JRC, Farndon JR, Needham GK, Malcolm AJ and Harris AL. (1987). Lancet, i, 1398 ± 1402. Shin DM, Ro JY, Hong WK and Hittelman WN. (1994). Cancer Res., 54, 3153 ± 3159. Theuer CP, Fitzgerald DH and Pastan I. (1993). J. Urol., 149, 1626 ± 1632. Todd R, Dono BR, Gertz R, Chang AL, Chow P, Matossian K, McBride J, Chiang T, Gallagher GT and Wong DT. (1989). Carcinogenesis, 10, 1553 ± 1556. Toi M, Osaki A, Yamada H and Toge T. (1991). Eur. J. Cancer, 27, 977 ± 980. Veale D, Ashcroft T, Marsh C, Gibson GJ and Harris AL. (1987). Br. J. Cancer, 55, 513 ± 516. Weichselbaum RR, Dunphy EJ, Beckett MA, Tybor AG, Moran WJ, Goldman ME, Vokes EE and Panje WR. (1989). Head & Neck, 11, 437 ± 442. Yamamoto T, Kamata N, Kawano H, Shimizu S, Kuroki T, Toyoshima K, Rikimaru K, Nomura N, Ishizaki R, Pastan I, Gamou S and Shimizu N. (1986). Cancer Res., 46, 414 ± 416.