Association of epidermal growth factor receptor (EGFR) gene mutations with EGFR amplification in advanced non‐small cell lung cancer

Abstract

Somatic mutations in the epidermal growth factor receptor (EGFR) gene are associated with the response to EGFR tyrosine kinase inhibitors in patients with non‐small cell lung cancer (NSCLC). Increased EGFR copy number has also been associated with sensitivity to these drugs. However, given that it is often difficult to obtain sufficient amounts of tumor tissue for genetic analysis from patients with advanced NSCLC, the relationship between these two types of EGFR alterations has remained unclear. We have now evaluated EGFR mutation status both by direct sequencing and with a high‐sensitivity assay, the Scorpion‐amplification‐refractory mutation system, and have determined EGFR copy number by fluorescence in situ hybridization (FISH) analysis in paired tumor specimens obtained from 100 consecutive patients with advanced NSCLC treated with chemotherapy. EGFR mutations or FISH positivity (EGFR amplification or high polysomy) were apparent in 18% (18/100) and 32% (32/100) of patients, respectively. The Scorpion‐amplification‐refractory mutation system was more sensitive than direct sequencing for the detection of EGFR mutations. Furthermore, EGFR mutations were associated with EGFR amplification (P = 0.009) but not with FISH positivity (P = 0.266). Our results therefore suggest the existence of a significant association between EGFR mutation and EGFR amplification in patients with advanced NSCLC. (Cancer Sci 2008; 99: 2455–2460)

The epidermal growth factor receptor (EGFR) is a receptor tyrosine kinase of the ErbB family and has been implicated in the proliferation and survival of cancer cells. Aberrant expression of EGFR has been detected in many human epithelial malignancies, including non‐small cell lung cancer (NSCLC).^{( 1} , ^{2 )} This receptor has therefore been identified as a promising target for anticancer therapy, and several agents have been synthesized that inhibit its tyrosine kinase activity. EGFR tyrosine kinase inhibitors (TKI) have been evaluated most extensively in individuals with NSCLC, and they have had a substantial impact on the treatment of this disease by offering additional therapeutic options for patients with advanced NSCLC.^{( 3} , ⁴ , ⁵ , ^{6 )}

Somatic mutations in the tyrosine kinase domain of EGFR have been detected in a subset of NSCLC patients who respond to EGFR TKI^{( 7} , ⁸ , ^{9 )} and have been shown to be closely associated with sensitivity to these drugs.^{( 10} , ¹¹ , ¹² , ¹³ , ^{14 )} Indeed, we and others have prospectively demonstrated a high response rate to EGFR TKI therapy in NSCLC patients with EGFR mutations.^{( 15} , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ^{21 )} An increased copy number of the EGFR gene, as revealed by fluorescence in situ hybridization (FISH), has also emerged as an effective molecular marker of EGFR TKI sensitivity in NSCLC.^{( 22} , ²³ , ^{24 )} We previously showed that EGFR mutation and EGFR amplification are associated in human NSCLC cell lines and that endogenous EGFR expressed in such cell lines positive for both of these EGFR alterations are activated constitutively.^{( 25 )} However, the relationship between EGFR mutation and FISH positivity for EGFR, which reflects gene amplification or high polysomy, has remained unclear.^{( 22} , ²³ , ²⁴ , ²⁶ , ^{27 )} Indeed, only a few studies have evaluated the relationship between mutation and gene copy number for EGFR because of the difficulty in obtaining tumor samples suitable for genetic analysis from individuals with advanced NSCLC. We previously showed that the Scorpion‐amplification‐refractory mutation system (ARMS) is a sensitive technique for the detection of EGFR mutations in tumor specimens such as pleural effusion fluid or tissue obtained by transbronchial needle aspiration.^{( 28} , ²⁹ , ^{30 )} In the present study, we evaluated EGFR mutation status in small tumor specimens from patients with advanced NSCLC both by direct sequencing and by Scorpion‐ARMS and compared the sensitivity of these methods for the detection of EGFR mutations. Furthermore, we determined EGFR copy number by FISH analysis in paired tumor specimens and examined its relationship to EGFR mutation.

Materials and Methods

Patients.  The present retrospective study recruited consecutive patients with advanced NSCLC who received chemotherapy at Kinki University Hospital between January 2003 and December 2005. Patients eligible for the study had histologically confirmed stage III or IV NSCLC that was not curable by surgical resection or radiotherapy, irrespective of the presence of measurable lesions or good performance status (PS). Patients with recurrence after surgical resection were excluded. Complete clinical information and tissue blocks suitable for genetic analysis were available for 100 patients. We examined the relationship between EGFR mutation and EGFR copy number as well as the influence of these EGFR alterations on clinical outcome. Tumor response was assessed by computed tomography and evaluated according to the Response Evaluation Criteria in Solid Tumors.^{( 31 )} Survival was calculated from the date of initiation of chemotherapy either to the date of death from any cause or to the date of last contact. Some patients had been receiving EGFR TKI treatment before the demonstration in 2004 that mutations in EGFR confer increased sensitivity to these drugs. Moreover, many patients had already died before the initiation of our genetic analysis, preventing us from obtaining informed consent. The institutional review board therefore approved our study protocol with the conditions that samples would be processed anonymously and analyzed only for somatic mutations (not for germline mutations) and that the study would be disclosed publicly, according to the Ethical Guidelines for Human Genome Research published by the Ministry of Education, Culture, Sports, Science, and Technology, the Ministry of Health, Labor, and Welfare, and the Ministry of Economy, Trade, and Industry of Japan. The present study also conforms to the provisions of the Declaration of Helsinki.

Identification of EGFR mutations.  The tumor specimens were fixed with formalin and embedded in paraffin. DNA was extracted with the use of a QIAamp Micro kit (Qiagen K.K., Tokyo, Japan) from tumor tissue derived either by macrodissection or by laser‐capture microdissection carried out to enrich tumor cells. Polymerase chain reaction‐based direct sequencing of exons 18–21 and ARMS with designed ‘Scorpion’ primers were applied for the allele‐specific detection of EGFR mutations. Only the following previously described mutations^{( 7} , ^{8 )} were classified as mutations in the present study: G719X in exon 18, deletion of E746 to A750 or of neighboring residues in exon 19, as well as L858R and L861Q in exon 21. Patients were regarded as EGFR mutation positive if a mutation in EGFR was detected either by direct sequencing or by ARMS. All mutations were confirmed by analysis of at least two independent amplification products.

Determination of EGFR copy number.  EGFR copy number was determined by FISH analysis with the use of dual‐color DNA probes (LSI EGFR SpectrumOrange/CEP 7 SpectrumGreen; Vysis, Downers Grove, IL, USA). The tumor specimens were classified into six categories on the basis of the FISH results, as described previously.^{( 22 )} Those with high polysomy (≥4 copies of EGFR in ≥40% of cells) or gene amplification (presence of a tight EGFR gene cluster and a ratio of EGFR to chromosome 7 of ≥2 or ≥15 copies of EGFR per cell in ≥10% of cells analyzed) were considered FISH positive, with those in the remaining categories being considered FISH negative.

Statistical analysis.  The relationships among EGFR status, clinical characteristics, and tumor response to EGFR TKI were analyzed with Fisher's exact test as appropriate. Survival curves were constructed by the Kaplan–Meier method, and the differences in survival between patient subgroups were compared by the log‐rank test. The impact of various factors on survival was evaluated by univariate and multivariate analysis according to the Cox regression model. A P‐value <0.05 was considered statistically significant. All statistical analysis was carried out with StatView software (SAS Institute, Cary, NC, USA).

Results

Patient characteristics.  Between January 2003 and December 2005, a total of 125 consecutive patients diagnosed histologically with advanced NSCLC underwent chemotherapy at Kinki University Hospital. Tissue specimens from 100 patients were assessable for both EGFR mutation and EGFR copy number. Of these specimens, 72 were obtained by bronchoscopic biopsy, 15 by percutaneous needle biopsy (12 from lung, two from bone, and one from lymph node), six by thoracoscopic biopsy, and seven by surgery for diagnosis or palliative therapy. The clinical characteristics of these 100 patients are shown in Table 1. Most of the patients were male (64%) and had a history of smoking (67%), and adenocarcinoma was the most prevalent tumor histology (61%). Most patients (83%) also had a good Eastern Cooperative Oncology Group PS (0 or 1), and 63% received second‐line or subsequent rounds of chemotherapy. Fifty‐three patients (53%) were treated with EGFR TKI. Seventy patients (70%) had died by the time of genetic analysis, with the median follow‐up time for the 30 survivors being 14.6 months.

Table 1.

Characteristics of patients with advanced non‐small cell lung cancer (n = 100)

Characteristic	Subset	No. patients
Sex	Male	64
	Female	36
Smoking history	Never‐smoker	33
	Smoker	67
Tumor histology	Adenocarcinoma	61
	Other	39
Eastern Cooperative Oncology Group performance status	0	24
	1	59
	≥2	17
No. chemotherapies	1	37
	≥2	63

EGFR alterations in non‐small cell lung cancer.  Patients were analyzed for EGFR mutations by direct sequencing of exons 18 through 21 and by Scorpion‐ARMS (Table 2). EGFR mutations, consisting of in‐frame deletions in exon 19 (n = 5) and point mutations in exon 21 (n = 13), were detected in 18 patients (18%). Eight EGFR mutations were detected by direct sequencing and 16 mutations were detected by Scorpion‐ARMS. Ten of the 16 mutations detected by Scorpion‐ARMS were not identified by direct sequencing. However, two of the deletions in exon 19 (E746_S752 and E746_T751) that were detected by direct sequencing were not identified by Scorpion‐ARMS, given that the Scorpion primers were designed only for detection of the E746_A750 deletion in exon 19. EGFR mutations were significantly more frequent in tumors of women than in those of men (33 vs 9%), in adenocarcinomas than in tumors with other histologies (28 vs 3%), and in never‐smokers than in smokers (42 vs 6%) (Fig. 1a). One of the 18 EGFR mutations was detected in a squamous cell carcinoma. Determination of EGFR copy number by FISH analysis revealed gene amplification in six patients and high polysomy in 26 patients, with 32 patients thus being classified as FISH positive (Table 3). In contrast to EGFR mutation, FISH positivity was not associated with sex, tumor histology, or smoking status (Fig. 1b). Although no relationship was apparent between EGFR mutation and FISH positivity (gene amplification or high polysomy), EGFR mutation and EGFR amplification were significantly associated (Table 4). The clinicopathological and genetic features of patients with EGFR mutations are shown in Table 5.

Table 2.

Detection of epidermal growth factor receptor (EGFR) mutations by direct sequencing or amplification‐refractory mutation system (ARMS) (n = 100)

Site	Mutation	Direct sequencing	ARMS	Direct sequencing or ARMS
Exon 19	15‐bp deletion	1	3	3
	16‐bp deletion	1	0	1
	19‐bp deletion	1	0	1
Exon 21	L858R	5	13	13
Total		8 (8%)	16 (16%)	18 (18%)

Figure 1.

Sex, tumor histology, and smoking status of patients with advanced non‐small cell lung cancer and with either (a) epidermal growth factor receptor (EGFR) mutations or (b) a high EGFR copy number. Ad, adenocarcinoma. *P‐values were determined by Fisher's exact test.

Table 3.

Determination of epidermal growth factor receptor gene copy number by fluorescence in situ hybridization (FISH) analysis (n = 100)

FISH status	Finding	No. patients
Positive	Gene amplification	6
	High polysomy	26
	Total	32
Negative	Low polysomy	35
	High trisomy	2
	Low trisomy	26
	Disomy	5
	Total	68

Table 4.

Relationship between epidermal growth factor receptor (EGFR) mutation and either fluorescence in situ hybridization (FISH) status of EGFR amplification

Mutation status	FISH status		Gene amplification
	Positive	Negative	Positive	Negative
Positive (n = 18)	8	10	4	14
Negative (n = 82)	24	58	2	80
P‐value*		0.266		0.009

^* Determined by Fisher's exact test.

Table 5.

Clinicopathological and genetic features of patients with epidermal growth factor receptor (EGFR) mutations

No.	Age (years)	Sex	Smoking status	Histology	Response to EGFR TKI	Type of EGFR mutation		EGFR copy number
						Sequencing	ARMS
1	72	F	Never	Ad	PR		L858R	Low trisomy
2	58	F	Never	Ad	PR	L858R	L858R	Gene amplification
3	81	F	Never	Ad	SD	L858R	L858R	High polysomy
4	72	F	Never	Ad	NE		L858R	Gene amplification
5	48	M	Smoker	Ad	SD		L858R	Low trisomy
6	67	F	Never	Ad	SD		L858R	Low trisomy
7	59	F	Never	Ad	PR		L858R	High polysomy
8	78	M	Smoker	Ad			L858R	High trisomy
9	71	F	Never	Ad	PR		L858R	Low polysomy
10	82	F	Never	Ad	PR	L858R	L858R	Low trisomy
11	67	F	Never	Ad		L858R	L858R	High polysomy
12	87	F	Never	Sq	PR	L858R	L858R	Low polysomy
13	78	M	Never	Ad			L858R	Gene amplification
14	56	F	Never	Ad	PR		(E746_A750)del	Low polysomy
15	63	M	Never	Ad	PD	(E746_A750)del	(E746_A750)del	Gene amplification
16	63	M	Smoker	Ad	PR		(E746_A750)del	Low polysomy
17	61	M	Smoker	Ad	PR	(E746_S752)del insV		Low trisomy
18	73	F	Never	Ad	PR	(E746_T751)del insS		High polysomy

Ad, adenocarcinoma; ARMS, amplification‐refractory mutation system; NE, not evaluated; PD, progressive disease; PR, partial response; SD, stable disease; Sq, squamous cell carcinoma; TKI, tyrosine kinase inhibitor.

Overall survival.  For the total patient population, the median overall survival was 12.3 months, with a 1‐year survival rate of 51.7%. Univariate analysis revealed that overall survival was significantly longer in women, never‐smokers, patients with a favorable PS, and those with EGFR mutations (Table 6; Fig. 2a). In contrast, no difference in overall survival was apparent between FISH‐positive and FISH‐negative patients (Table 6; Fig. 2b). We also carried out multivariate analysis to identify factors that contribute to overall survival, with covariates including clinicopathological and genetic factors (sex, smoking history, tumor histology, PS, EGFR mutation status, FISH status). Female sex and favorable PS were found to be independent prognostic factors (Table 6).

Table 6.

Univariate and multivariate analyses of prognostic factors for overall survival

Factor	Univariate analysis			Multivariate analysis
	HR	95% CI	P‐value	HR	95% CI	P‐value
Sex (female/male)	0.54	0.32–0.91	0.021	0.55	0.32–0.93	0.025
Smoking history (never‐smoker/smoker)	0.50	0.30–0.85	0.011
Histology (adenocarcinoma/other)	0.64	0.39–1.05	0.077	0.68	0.40–1.14	0.141
ECOG PS (0/≥1)	0.44	0.24–0.79	0.006	0.48	0.29–0.86	0.019
EGFR mutation status (positive/negative)	0.52	0.28–0.97	0.039
FISH status (positive/negative)	1.36	0.82–2.23	0.231	1.49	0.88–2.50	0.130

CI, confidence interval; ECOG, Eastern Cooperative Oncology Group; EGFR, epidermal growth factor receptor; FISH, fluorescence in situ hybridization; HR, hazard ratio; PS, performance status. Multivariate analysis was carried out using the stepwise method (include, <0.05; exclude, >0.2). Significant P‐values are shown in bold.

Figure 2.

Kaplan–Meier plots of overall survival in patients with advanced non‐small cell lung cancer and either (a) with or without epidermal growth factor receptor (EGFR) mutations or (b) with or without a high EGFR copy number. FISH, fluorescence in situ hybridization.

Responsiveness to epidermal growth factor receptor tyrosine kinase inhibitor treatment.  Of the 53 patients treated with EGFR TKI, 40 individuals were assessable for objective response. Whereas the rate of response to EGFR TKI treatment for patients with EGFR mutations was significantly higher than that for those without such mutations (71.4 vs 11.5%, P < 0.001), there was no significant association between FISH status and responsiveness to EGFR TKI (44.4 vs 29.0% for FISH‐positive vs FISH‐negative patients, respectively, P = 0.437).

Discussion

We have analyzed both EGFR mutation and EGFR copy number in paired tumor specimens as well as the relationship between these two types of EGFR alterations in advanced NSCLC. We used two methods to detect EGFR mutations, direct sequencing and Scorpion‐ARMS, which identified eight and 16 mutations, respectively. Direct sequencing failed to detect 10 of the 16 mutations identified by Scorpion‐ARMS. Of the 10 patients with EGFR mutations detected by Scorpion‐ARMS alone, seven were assessable for an objective response to EGFR TKI, with five exhibiting a partial response and two having stable disease. Consistent with previous observations,^{( 28} , ²⁹ , ^{30 )} our data thus indicate that Scorpion‐ARMS is more sensitive than direct sequencing for detection of the two major types of EGFR mutation that reflect responsiveness to EGFR TKI. It should be noted, however, that most polymerase chain reaction‐based systems for mutation analysis, including Scorpion‐ARMS, are able to detect only known EGFR mutations targeted by the designed primers. Indeed, two minor variants of deletion mutation in exon 19 were not identified by Scorpion‐ARMS in the present study. Given the exclusion of recurrence after surgical resection in our study, most tumor specimens analyzed were obtained either by transbronchial lung biopsy or by percutaneous needle lung biopsy. The amount of tumor tissue obtained by these procedures is limited, but our results suggest that it is sufficient both for histopathological analysis and for the detection of EGFR mutations by Scorpion‐ARMS in patients with advanced NSCLC.

Scorpion‐ARMS identified three E746_A750 deletion mutations in exon 19 and 13 L858R point mutations in exon 21 in the present study. The frequency of the E746_A750 mutation detected by Scorpion‐ARMS thus appeared low compared with that of the L858R mutation. Previous studies have shown that the incidence of the E746_A750 deletion is approximately the same as that of the L858R mutation.^{( 10} , ^{12 )} The sensitivity of Scorpion‐ARMS for detection of the E746_A750 deletion is equivalent to that for detection of the L858R point mutation. The low frequency of the E746_A750 deletion mutation in the present study is thus likely due to the small number of samples.

Previous studies have revealed a higher prevalence of EGFR mutations in East Asians than in Caucasians.^{( 4} , ¹⁰ , ¹¹ , ¹² , ²⁰ , ²² , ²⁴ , ²⁶ , ²⁷ , ³² , ³³ , ³⁴ , ³⁵ , ^{36 )} The prevalence of EGFR mutations in our Japanese cohort was low (18%) compared with values determined previously for East Asian populations. Given that most previous studies examined only individuals treated with EGFR TKI, patient selection based on clinical predictors might have led to an increase in the proportion of subjects with adenocarcinoma histology, a factor known to be associated with EGFR mutations. In contrast, our study was carried out with consecutive cases irrespective of EGFR TKI treatment. The relatively low proportion of patients with adenocarcinoma histology (61%) in our cohort is therefore consistent with the low prevalence of EGFR mutations. However, the FISH positivity of 32% in our study is similar to that in previous studies that adopted the same criteria, with values ranging from 31 to 48%.^{( 22} , ²³ , ²⁴ , ²⁶ , ^{27 )} Consistent with previous results,^{( 1} , ⁷ , ⁸ , ⁹ , ^{12 )} EGFR mutations were significantly more frequent among women, never‐smokers, and patients with adenocarcinoma in the present study. In contrast, neither EGFR amplification (analysis not shown) nor FISH positivity was associated with any such clinicopathological factor in our study, although the relationship between EGFR amplification and never‐smoking status approached statistical significance (P = 0.090).

The relationship between EGFR mutation and FISH positivity (gene amplification or high polysomy) in NSCLC patients has remained unclear.^{( 22} , ²³ , ²⁴ , ²⁶ , ^{27 )} In the present study, we have demonstrated a significant relationship between EGFR mutation and EGFR amplification, but not between EGFR mutation and FISH positivity, in tumor specimens from patients with advanced NSCLC. EGFR mutant alleles were previously found to be amplified selectively, resulting in a high EGFR copy number, as detected by quantitative real‐time polymerase chain reaction analysis.^{( 12 )} EGFR amplification has also been shown to be acquired during invasive growth of lung adenocarcinoma with EGFR mutations.^{( 37 )} Furthermore, recent studies have found that an increase in EGFR copy number is a relatively late event in NSCLC pathogenesis^{( 38 )} and that EGFR mutation precedes EGFR amplification but not necessarily high polysomy.^{( 37} , ^{39 )} These observations thus support the existence of a close association between EGFR mutation and EGFR amplification. We previously showed that EGFR mutation was significantly associated with EGFR amplification in human NSCLC cell lines and that endogenous EGFR expressed in such cell lines that manifested both of these EGFR alterations were activated constitutively as a result of ligand‐independent dimerization.^{( 25 )} However, the biological consequences of high polysomy for EGFR have not been elucidated. We did not find any cut‐off value of high polysomy that was associated with EGFR mutation. We therefore propose that EGFR amplification, but not high polysomy, plays a key role in the pathogenesis of NSCLC and correlates with EGFR mutation.

We sought to determine whether EGFR mutation or EGFR copy number might affect overall survival of NSCLC patients. Previous studies of EGFR TKI have suggested that EGFR mutation is a favorable prognostic indicator for patients with NSCLC.^{( 35} , ^{36 )} We also found that the survival time of patients with EGFR mutations was longer than that of those without them (18.0 vs 11.6 months, P = 0.036) in the univariate analysis. However, interpretation of this result requires that the effect of EGFR TKI on survival be taken into account, given that 83% (15/18) of patients with EGFR mutations were treated with EGFR TKI compared with only 46% (38/82) of those without such mutations. Indeed, analysis of survival after initiation of EGFR TKI treatment as a second‐line or subsequent therapy revealed a survival time of 15.6 months for mutation‐positive patients vs 6.0 months for mutation‐negative patients in our study. It was therefore not possible to determine the prognostic significance of EGFR mutation for NSCLC patients. To clarify whether EGFR mutation is a predictor of sensitivity to EGFR TKI or a prognostic indicator for NSCLC patients, we are currently carrying out a phase III randomized study comparing platinum‐based chemotherapy with gefitinib in chemotherapy‐naive NSCLC patients with EGFR mutations. Patients with FISH‐positive tumors tended to have a shorter survival time than did those with FISH‐negative tumors (10.7 vs 13.8 months), although this difference was not statistically significant. This result is consistent with previous observations indicative of an association between high EGFR copy number and poor prognosis for certain malignancies, including NSCLC.^{( 1} , ^{40 )}

In conclusion, we have analyzed both EGFR mutation and EGFR copy number in paired tumor specimens from patients with advanced NSCLC. We found that Scorpion‐ARMS is more sensitive than direct sequencing for detection of EGFR mutations in small tumor specimens. Furthermore, we showed that EGFR mutation was significantly associated with EGFR amplification but not with FISH positivity. These observations warrant confirmation in further studies as well as exploration of the biological mechanisms of the relationship between EGFR mutation and EGFR amplification. The effects of EGFR mutation and EGFR copy number on clinical outcome in individuals with advanced NSCLC also warrant investigation in a prospective study.