This article is mentioned in:

Introduction

Epidermal growth factor receptor (EGFR) is a 170-kDa transmembrane glycoprotein receptor that exhibits tyrosine kinase activity, which regulates cell growth. EGFR expression was frequently observed in squamous cell carcinomas of the head and neck (SCCHN) (1). Since EGFR signaling was found to control not only cell growth, but also angiogenesis and DNA repair (2,3), it may be a molecular target in patients with SCCHN (4,5).

Gefitinib (ZD1839, Iressa®), an anilinoquinazoline-based inhibitor specific for EGFR tyrosine kinase, was shown to have clinical efficacy in patients with non-small cell lung carcinoma (NSCLC) (1). Since mutations in the tyrosine kinase domain of the EGFR gene were reported to increase the binding affinity of gefitinib, the presence of EGFR mutations in this domain is predictive of the clinical benefits of gefitinib treatment for NSCLC patients (6,7). Gefitinib has also shown clinical efficacy in some patients without EGFR mutations (4,5). Thus, in such patients, EGFR gene amplifications are critical for gefitinib efficacy (6,7). Previous studies have shown that mutations in the kinase domain of the EGFR gene occur less frequently in SCCHN than in NSCLC (1,4). The clinical efficacy of gefitinib is due not only to its direct action on EGFR tyrosine kinase, but on its indirect action through activation of the immune system (9). For example, to demonstrate the direct action of EGFR tyrosine kinase, it was shown that SCCHN cell lines with heterozygous mutations (G/G→G/A) at nucleotide 2607 of EGFR are more sensitive to gefitinib than wild-type EGFR (4). To further investigate these observations, the mechanism by which the G/A mutation confers sensitivity to gefitinib was examined.

Materials and methods

Cell lines and cell culture

The 16 human SCCHN cell lines and their sites of origin included YCU-M862, YCU-M911 and KCC-M871, from tumors of the mesopharynx; YCU-H891, from a tumor of the hypopharynx; KCC-L871 and YCU-L891, from tumors of the larynx; KCC-T871, YCU-T891, YCU-T892 and KCC-T873, from tumors of the tongue; KCC-TCM902, KCC-TCM903 and KCC-TCM901, from metastatic tumors of 3 tongue carcinomas; KCC-OR891, from tumors of the oral floor; and KCC-MS871 and YCU-MS861, from tumors of the maxillary sinus. The cell lines were maintained in RPMI-1640 supplemented with 10% fetal bovine serum and cultured at 37°C in a humidified atmosphere of 5% CO2/95% air.

Fluorescence in situ hybridization (FISH)

The cells were fixed in carunor solution, placed on glass slides and hybridized with LSI EGFR Dual Color Probe (Abbott Molecular, CA, USA) (1,2). The signals were counted for 60 cells in each cell line and data were expressed as mean ± SD.

mRNA half life

Each cell line was treated with 5 μg/ml actinomycin D (Wako Pure Chemical Industries, Osaka, Japan) and total RNA was purified using Isogen RNA purification kit (Nippon Gene, Tokyo, Japan). Total RNA was then reverse-transcribed to cDNA using avian myeloblastosis virus (AMV) reverse transcriptase (Takara, Tokyo, Japan) and amplified by real-time quantitative PCR (Stratagene® MX3000P; Agilent Technologies, Santa Clara, CA, USA) with Stratagene Brilliant II Fast SYBR® Green QPCR Master Mix (Agilent Technologies). The specific primer sets are shown in Table I. The amplification protocol consisted of 30 cycles of denaturation at 95°C for 30 sec, annealing at 55°C for 1 min and extension at 72°C for 30 sec.

Table I QPCR primer sets.

Table I

QPCR primer sets.

Genes	Sequence (5′→3′)
EGFR	Forward: TTCGATGATCAACTCACGGAAC
	Reverse: GCCACCCATATGTACCATCGAT
β-actin	Forward: CAGCTGAGAGGGAAATCGTG
	Reverse: GCTGGTTGCCAATAGTGAATG

Statistical analysis

The groups were compared using paired Student’s t-tests and Chi-square tests. Linear regression analysis was used to compare the EGFR mRNA half life with the EGFR gene copy number, the steady state levels of EGFR mRNA and the EGFR protein levels. P<0.05 was considered to be statistically significant.

Results

EGFR copy number

The 16 SCCHN cell lines were divided into two groups: those heterozygous for the mutation (G/A) at nucleotide 2607 of the EGFR gene (9 cell lines) and those with the wild-type (G/G) sequence (7 cell lines). The mean number of EGFR copies in the cell lines was 3.59±2.14 and 1.58±0.32, respectively (Fig. 1, Table II) (4), or 2.27-fold higher in cell lines with the mutant sequence. It was found that 10 of the 16 cell lines had ≥2 copies of the EGFR gene, whereas the other 6 had <2 copies. The Chi-square test showed a statistically significant correlation between the G/A mutation and EGFR copy number (Table III). The EGFR copy number was significantly higher in YCU-OR891 than in the remaining cell lines.

Figure 1 Status of the EGFR gene in SCCHN cell lines carrying the G/A and G/G genotypes at codon 2607. (A and B) Number of copies of the EGFR gene, as determined by FISH. Representative results are shown. (A) YCU-MS861, 2.02 copies/cell. (B) KCC-TCM903, 1.13 copies/cell. (C and D) EGFR mRNA half-life, determined following the addition of actinomycin D (5 μg/ml). Representative results are shown. (C) KCC-T871, 6 h. (D) YCU-H891, 13 h. (E) Average mRNA half life in the seven cell lines with the G/G genotype and the nine cell lines with the G/A genotype. Error bars indicate SD. *P<0.05.

Table II A summary of EGFR protein, mRNA, gefitinib IC50 value, SNP, FISH copy number ratio and mRNA half life.

Table II

A summary of EGFR protein, mRNA, gefitinib IC50 value, SNP, FISH copy number ratio and mRNA half life.

Cell line	EGFR proteina	EGFR mRNA	IC50 valuea	Genotype at 2607b	FISH X/Y	mRNA half life (h)
YCU-MS861	0.15	0.20	66.3	G/G	1.13	4.3
YCU-M862	1.85	0.59	64.8	G/G	1.33	5.4
KCC-M871	0.46	0.10	52.1	G/G	1.47	3.9
KCC-L871	0.58	0.04	58.0	G/G	1.54	5.6
YCU-T892	2.25	4.43	64.8	G/G	1.68	3.0
YCU-M911	0.95	0.55	61.4	G/G	1.82	5.2
KCC-T873	2.96	0.85	85.0	G/G	2.12	5.6
KCC-TCM903	0.38	0.25	51.4	G/A	2.02	6.0
KCC-T871	0.60	0.09	44.4	G/A	2.55	6.0
KCC-TCM901	0.77	0.25	43.9	G/A	2.80	5.8
YCU-H891	0.21	2.38	69.0	G/A	2.89	13.0
YCU-L891	0.91	1.14	52.2	G/A	2.89	8.5
YCU-T891	0.56	2.46	56.0	G/A	2.89	18.0
KCC-MS871	0.23	1.35	41.6	G/A	3.53	15.0
KCC-TCM902	0.25	1.16	50.2	G/A	3.60	11.2
YCU-OR891	0.47	0.16	53.6	G/A	9.16	8.5
Average	0.85	1.00	57.2		2.71	7.8

a Data from Taguchi et al (1).

b SNP, single nucleotide polymorphisms.

Table III Heterozygous EGFR mutants contain a high copy number of EGFR.

Table III

Heterozygous EGFR mutants contain a high copy number of EGFR.

Copy no.	G/G	G/A
≥2	1	9
<2	6	0

[i] Chi-square test, p=0.00044.

Association between EGFR copy number and mRNA half life

Real-time qPCR analysis revealed that the EGFR gene copy number was positively correlated with the steady state levels of EGFR mRNA and the EGFR mRNA half life (Fig. 2C). Notably, the mRNA half life was inversely correlated with EGFR protein concentration (Fig. 2A).

Figure 2 Linear regression analysis between EGFR mRNA half life and (A) EGFR gene copy number, (B) the steady state levels of EGFR mRNA and (C) EGFR protein levels. EGFR mRNA levels were determined following cell culture with actinomycin D (5 μg/ml). The EGFR protein levels were from our previously published results (4).

Discussion

The efficacy of gefitinib in patients with NSCLC was thought to depend on specific EGFR gene mutations as opposed to the concentration of EGFR protein. In particular, mutations in the tyrosine kinase domain, consisting of exons 18–21, were regarded as critical for a response to gefitinib. Dominant mutations or deletions of 2–15 nucleotides between codons 740 and 753 in exon 19 were found to increase the tyrosine kinase inhibitory activity of gefitinib due to conformational changes at the ATP-binding site (4). We found that a number of SCCHN cell lines have the G/A genotype, consisting of a heterozygous G→A transition at nucleotide 2607 in exon 20 of the EGFR gene (4). Cell lines with the G/A genotype were significantly more sensitive to gefitinib (lower IC50 values) than cell lines with the wild-type G/G genotype (P=0.016). We then focused on the relationship between gene expression and sensitivity to gefitinib in vitro. The FISH assay showed that cell lines carrying the G/A mutation have a significantly higher EGFR copy number than wild-type cell lines. Moreover, the steady state levels of EGFR mRNA were higher in cells with the G/A genotype than in those with the G/G genotype, suggesting a close correlation between the EGFR copy number and the steady state levels of EGFR mRNA. Notably, concentrations of EGFR protein were inversely correlated with the EGFR mRNA half life and steady state levels.

The G/A mutation proved to be synonymous, which may be due to impaired translation as opposed to protein instability. Generally, synonymous mutations are thought to be silent. However, cells with the G/A mutation produce less EGFR protein than those with the G/G wild-type, thereby suggesting that the mutation affected the translation rate. A previous study showed that a synonymous GAA to GAG mutation resulted in a 3-fold difference in the rate of translation due to the binding properties of tRNA to each codon (10). Moreover, a silent synonymous mutation in the MDR1 (multidrug resistance 1)/ABCB1 gene, which encodes P-glycoprotein, does not alter the mRNA or protein levels, but affects protein conformation, thereby altering the interaction between the substrate and inhibitor (11). Thus, synonymous mutations may affect mRNA structure and stability, the kinetics of translation and alternative splicing (12). We found that the heterozygous G/G→G/A mutation at codon 2607 of the EGFR gene was correlated with gene amplification, prolongation of EGFR mRNA half life and an increase in the steady state concentrations of EGFR mRNA. Additionally, the mutation was inversely correlated with the EGFR protein level, suggesting that this mutation affects translation efficacy. Although cell lines carrying the G/A mutation were more sensitive to gefitinib, we did not observe a significant correlation between gefitinib IC50 and the EGFR molecular status, including the copy number and the mRNA half life, due to the narrow range of IC50 values in a limited number of cell lines. However, combined with our previous findings, which demonstrated that G/A mutants exhibited a higher sensitivity to gefitinib, the down-regulation of EGFR protein expression, possibly due to a reduced translation efficacy, may be closely associated with gefitinib efficacy. Since the efficacy of gefitinib is thought to be independent of the EGFR protein concentration, but dependent on EGFR mutations in the tyrosine kinase domain, the structure of EGFR protein may have been altered by the synonymous mutation. Further studies are required to clearly determine the relationship between the efficacy of gefitinib and the protein structure of G/A mutants of the EGFR protein.

References