EGFR Mutation Status in Japanese Lung Cancer Patients: Genotyping Analysis Using LightCycler

Lung cancer is a major cause of death from malignant diseases because of its high incidence, malignant behavior, and lack of major advancements in treatment strategy (1). Lung cancer was the leading indication for respiratory surgery (42.2%) in 1998 in Japan (2). More than 15,000 patients underwent surgical operation at Japanese institutions in 1998 (2). The clinical behavior of the lung cancer is largely associated with its stage. The cure of the disease by surgery is only achieved in cases representing an early stage of lung cancer (3).

The epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, gefitinib, has been approved in Japan for the treatment of non–small cell lung cancer from 2002. Although EGFR is more abundantly expressed in lung carcinoma (4, 5), EGFR expression, as detected by immunohistochemistry, did not reveal any obvious relationship with response to gefitinib (6). Clinical trial have revealed significant variability in the response to gefitinib, with higher response in Japanese patients than in predominantly European-derived population (27.5% versus 10.4%; ref. 7). The partial clinical responses to gefitinib have been observed most frequently in women, in nonsmokers, and in patients with adenocarcinoma (8–10). More recently, we have collaborated with Dana-Farber Cancer Institute and found that novel EGFR mutations status at ATP binding pockets in Japanese non–small cell lung cancer patients were correlated with the clinicopathologic features related to good response to gefitinib (11). Actually, EGFR mutations in lung cancer have been correlated with clinical response to gefitinib therapy in vivo and in vitro (11–13).

The standard for experimental detection of mutations is direct sequencing of DNA samples from the tissues. For known mutations, real-time polymerase chain detection followed by melting curve analysis, using hybridization probes, is highly sensitive, rapid, and an efficient alternative approach to mutation detection (14–16).

To determine the EGFR mutation status in Japanese lung carcinoma for screening and diagnostic purposes, we wanted to develop a faster and easy method to detect EGFR mutations. In this report, we investigated EGFR mutation status by the real-time reverse transcription-PCR assay using LightCycler (17) mutation-specific sensor and anchor probes. With this method, 32 samples were genotyped within 1 hour without the need of any post-PCR sample manipulation. The findings were compared with the clinicopathologic features of lung cancer.

Patients. The study group included 102 lung cancer patients who had undergone surgery (but did not receive gefitinib) at the Department of Surgery II, Nagoya City University Medical School, between 1997 and 2000. The study group also included 16 lung cancer patients who had undergone surgery at the Department of Surgery, National Hospital Organization, Kinki-chuo Chest Medical Center, and were subsequently treated with gefitinib. These 16 samples were sequenced by ABI prism 3100 analyzer (Applied Biosystems Japan, Ltd., Tokyo, Japan; data not shown) and analyzed by ABI prism Seq Scape version 2.1.1. The lung tumors were classified according to the general rule for clinical and pathologic record of lung cancer in Japan (18). All tumor samples were immediately frozen and stored at −80°C until assayed.

The clinical and pathologic characteristics of the 102 lung cancer patients are as follows: 52 cases at stage I, 16 at stage II, and 34 at stage III to IV. The mean age was 65.5 years (range, 42-85). Among the 102 lung cancer patients, 49 (48%) were diagnosed as having adenocarcinoma, 32 (31.4%) squamous cell carcinoma, 9 (8.8%) adenosquamous cell carcinoma, and 7 (6.9%) small cell carcinoma.

PCR assays for EGFR. The genomic DNA was extracted from lung cancer tissues and matched normal lymphocytes from the peripheral blood using the Wizard SV Genomic DNA purification system (Promega Corporation, Madison, WI). Initially, 58 DNA samples were also extracted from lung cancer tissues from Nagoya City University and sequenced as reported in our previous paper (11). These sets of DNA were used as a positive and negative control for genotyping. DNA concentration was determined by spectrophotometry and adjusted to a concentration of 50 ng/mL. We then used 1 μL of each DNA for LightCycler analyses. To ensure the fidelity of DNA extraction, all samples were subjected to PCR amplification with oligonucleotide primers specific for exon 18 of the EGFR gene and then digested by SacI enzyme. The primer sequences for EGFR gene in exon 18 were as follows: the forward primer, 5-TCCAAATGAGCTGGCAAGTG-3, and the reverse primer, 5-TCCCAAACACTCAGTGAAACAAA-3 (397 bp). The cycling conditions were as follows: initial denaturation at 95°C for 15 minutes followed by 35 cycles at 95°C for 20 seconds, 57°C for 20 seconds, 72°C for 30 seconds, and one cycle of 72°C for 3 minutes. The products were purified by Qiagen PCR purification kit (Qiagen, Valencia, CA) and then digested with restriction enzyme at 37°C for 2 hours. The genotyping PCR reactions were done using LightCycler DNA Master Hybridization probes kit (Roche Molecular Biochemicals, Mannheim, Germany) in a 20 μL reaction volume. The primer sequences for EGFR gene in exon 18 were as follows: the forward primer, 5-TCCAATGAGCTGGCAAGTG-3, and the reverse primer, 5-TCCCAAACACTCAGTGAAACAAA-3 (397 bp). For the exon 18 genotyping, sensor (LC Red 640-GCACCGGAGCCCAGCA) and anchor (GCCAGGGACCTTATACACGTGCCGAA-Fluorescein) probes were used. The cycling conditions were as follows: initial denaturation at 95°C for 10 minutes, followed by 45 cycles at 95°C for 10 seconds, 60°C for 10 seconds, and 72°C for 16 seconds The primer sequences for EGFR gene in exon 19 were as follows: the forward primer, 5-CGTCTTCCTTCTCTCTCTGTC-3, and the reverse primer, 5-GACATGAGAAAAGGTGGGC-3 (175 bp). For the exon 19 genotyping, sensor (GCTATCAAAACATCTCC-Fluorescein) and anchor (LC Red 640-AAAGCCAACAAGGAAATCCTCGATGTGAGTTTCTGCTTTGCTGTGTGGGG) probes were used. The cycling conditions were as follows: initial denaturation at 95°C for 10 minutes followed by 45 cycles at 95°C for 10 seconds, 60°C for 10 seconds, and 72°C for 7 seconds. The primer sequences for EGFR gene in exon 21 were as follows: the forward primer, 5-GCTCAGAGCCTGGCATGAA-3, and the reverse primer, 5-CATCCTCCCCTGCATGTGT-3 (349 bp). The cycling conditions were as follows: initial denaturation at 95°C for 10 minutes, followed by 45 cycles at 95°C for 10 seconds, 57°C for 10 seconds, and 72°C for 14 seconds. For the exon 21 genotyping, sensor (Fluorescein-AGTTTGGCCCGCCCA) and anchor (LC Red 640-CCTCCTTCTGCATGGTATTCTTTCTCTTCCGCACCCAG) probes were used.

Statistical methods. Statistical analyses were done using the Mann-Whitney U test for unpaired samples and Wilcoxon's singed rank test for paired samples. Linear relationships between variables were determined by simple linear regression. Correlation coefficients were determined by rank correlation using Spearman's test and χ² test. The overall survival of lung cancer patients was examined by the Kaplan-Meier methods and differences were examined by the log-rank test, Breslow-Gehan-Wilcoxon test, and Cox proportional hazard regression model. All analyses were done using the StatView software package (Abacus Concepts, Inc., Berkeley, CA) and results were considered significant when P < 0.05.

Fidelity of allele-specific PCR confirmed by conventional PCR assay in lung cancer tissues. Using the exon 18 primer sets, a PCR product of 397 bp was obtained. We have analyzed the product using PCR-RFLP method. The wild-type DNA does not have a SacI site within the 397 bp. The PCR products digested with SacI were loaded with 2% agarose gel and wild-type DNA should be visualized as one band. However, if the substitution mutation G719S were present, the PCR products digested by SacI will be visualized as three bands. Using this method, PCR products were visualized from all lung cancer patients studied. In exon 18, a G719S mutation was found from one Nagoya specimen (stage Ia, well-differentiated adenocarcinoma with bronchioloalveolar carcinoma pattern at the edge of tumor, female, nonsmoker patient) and one Kinki specimen (Fig. 1A). These mutants were also analyzed by LightCycler. The anchor probe was matched for wild type. As shown in Fig. 1B for the G719S mutation in exon 18, the homozygous wild-type PCR product showed a single peak at 69°C, whereas the heterozygous products (mutant) showed an additional peak at 59°C. The LightCycler method using the mutation-specific probes confirmed the results with the restriction fragment analysis.

Genotyping of EGFR at exon 19 and exon 21 in lung cancer tissues. For exon 21 genotyping, the anchor probe was matched for L858R mutation. As shown in Fig. 2, for the L858R mutation in exon 21, the homozygous wild-type PCR product showed a single peak at 53°C, whereas the heterozygous products (mutant) showed an additional peak at 65°C. From the 102 lung cancer patients, 8 patients had the L858R mutation. One was male and seven were female. Seven were nonsmokers and one was a smoker (Brinkman index was 600). All eight patients had adenocarcinoma, one was moderately differentiated, and seven were well differentiated. Five of eight adenocarcinomas showed bronchioloalveolar carcinoma pattern at the edge of tumor. Thus, L858R mutation status was significantly correlated with gender, Brinkman index, pathologic subtypes, and differentiation of lung cancer (Table 1). Eight of eight PCR products from matched peripheral lymphocyte DNA showed a single peak, suggesting that the mutations were somatic. L858R mutation was also found in one nonsmoking female adenocarcinoma patient from Kinki-chuo Chest Medical Center.

For exon 19 genotyping, the anchor probe was matched for deletion type 1a (2,235-2,249 nucleotides deletion; deletion GGAATTAAGAGAAGC) mutation. As shown in Fig. 3, for the deletion 1a mutation in exon 19, the PCR product showed a single peak at 56°C, whereas the deletion 1b products (2,236-2,250 nucleotides deletion; deletion GAATTAAGAGAAGCA) showed a peak at 47°C. From the 102 lung cancer patients, seven patients had the deletion 1a mutation. Four were males and three were females. Three were nonsmokers and four were smokers. Four patients had adenocarcinoma, two had squamous cell carcinoma, and one had adenosquamous cell carcinoma. One of the tumors was moderately differentiated, two were poorly differentiated, and three were well differentiated. One of four adenocarcinomas showed bronchioloalveolar carcinoma pattern at the edge of tumor. Thus, deletion 1a mutation status was not significantly correlated with gender, Brinkman index, pathologic subtypes, and differentiation of lung cancer (Table 2). Five of seven PCR products from matched peripheral lymphocyte DNA were available and showed a single peak, suggesting that these mutations were somatic.

The mutations detected in lung cancer specimens from Kinki-chuo Chest Medical Center are summarized in Table 3. L858R mutation and deletion type 1a were found from partial response patients. On the other hand, G719S mutation was found from a patient with no response to gefitinib (progressive disease). A total of six mutations were found from 16 gefitinib-treated patients (37.5%). Taken together, 22 mutations were found from 117 examined samples in our analysis (18.8%).

The overall survival of 102 lung cancer patients from Nagoya City University, with follow-up through December 30, 2003, was studied in reference to the EGFR mutation status. There was no significant difference in the prognosis between the patients with wild-type EGFR (n = 86, 22 were dead) and the patients with mutation in the EGFR gene (n = 16, two were dead; log-rank test, P = 0.3608; Breslow-Gehan-Wilcoxon test, P = 0.4761), although the observation period was short.

We obtained findings that L858R EGFR mutation status was significantly correlated with gender, smoking history, and pathologic subtypes of lung cancers. This was in agreement with the recent reports that EGFR gene mutations are common in lung cancers from never smokers (13) and females with adenocarcinoma (11). Our analysis also suggested that the type of EGFR mutation might be correlated with the sensitivity of gefitinib therapy for lung cancers.

When the PCR is used for the detection of mutations in very small amounts of DNA, although we would like to start from biopsy samples in the future, it is usually necessary to use “nested PCR.” In this case, a DNA fragment is amplified with a first set of primers and part of the product is reamplified with a second set of primers complementary to sequences in the product. Recent developments in fast PCR and real-time detection of products make a more sensitive approach to detection of mutations possible (14–16, 19). We have optimized mutation detection, without nested PCR, using the LightCycler. This instrument measures fluorescence during PCR and can detect the SYBR Green dye when it is intercalated in double-stranded DNA, allowing the detection of double-stranded PCR product formation. The use of labeled probes homologous to the PCR product permits specific identification of PCR products (17). In the LightCycler, two adjacent probes were used, labeled with different fluorescent molecules. When the probes were bound to the single-stranded target, one to five bases apart, the 3′-end label of the 5′ probes came close to the 5′-end label of the 3′ probe, resulting in resonance and strong fluorescence at a specific wavelength. An advantage of this strategy is that hybridization of the probe is not restricted to the temperature range required for Taq polymerase to remove a base (19, 20). Further melting curves can be produced after PCR to assess the dissociation temperature of the probe. Mutations covered by the probe can be detected by a shift in melting temperature. The one-cycle analysis took ∼1 hour and could examine 32 samples.

Because so many EGFR mutation phenotypes were discovered, it would be of interest to determine whether resistance to EGFR inhibition emerges through secondary mutation as is the case in imatinib-treated chronic myelogenous leukemia (21). Our data showed that L858R mutation and deletion type 1a were found in gefitinib-sensitive patients; on the other hand, a G719S mutation was found in a gefitinib-resistant patient. Interestingly, recent data reported that L858R mutant (transfected cell) was inhibited at 10-fold lower concentrations of tyrosine kinase inhibitor; however, the deletion mutant seemed to have similar sensitivities as wild-type EGFR to drug (13). Thus, mutation phenotypes might be correlated with sensitivity for gefitinib therapy. Substitution mutation L858R is located adjacent to the highly conserved DFG motif in the activation motif. The activation loop was known to be important for autoregulation in many kinases (22). For example, the mutation in the activation loop of insulin receptor tyrosine kinase substantially increases the ability of the unphosphorylated kinase to bind ATP (23). From our data, this mutation pattern (L858R) might be more correlated with the populations, such as women, smoking, and adenocarcinoma.

DNA sequencing using the PCR methods described to date is time-consuming and, therefore, may not be suitable for a regular pretherapeutic screening program. Genechip technology is promising but still in its infancy, and adapting this technology to new polymorphisms is time-consuming and expensive. Real-time PCR, on the other hand, allows for easy adoption of new polymorphisms and possibly provides the best means for pretherapeutic genotyping in a clinical setting at present. We, therefore, developed three different PCRs to detect EGFR gene mutations and deletions. The advantages of real-time PCR are extensive. The faster PCR method and elimination of additional steps to analyze PCR products save time and minimize the risks of DNA contamination. Handling is facilitated and potentially toxic reagents, such as ethidium bromide stain, are avoided. We have only found 16 of 101 surgically removed samples from Nagoya City University and 6 of 16 gefitinib-treated samples from Kinki-chuo Chest Medical Center. Other mutations might have existed for these patients, although we have only checked the three most frequent mutations. The difference in the ratio of EGFR mutation between Nagoya and Kinki patients might have been caused by selection bias because gefitinib was known to be sensitive for female, nonsmoker, and adenocarcinoma patients. Actually, we have checked seven small cell carcinoma and three large cell carcinoma patients from Nagoya and no mutations were found from these patients.

Using the LightCycler reverse transcription-PCR assay described here, the determination of EGFR mutation status may be of clinical importance in predicting the sensitivity or resistance to gefitinib therapy for lung cancer. With this method, 32 samples were genotyped within 1 hour without the need of any post-PCR sample manipulation. Mutation detection using real-time PCR with hybridization probes and melting curve analysis can be used for the sensitive detection of DNA mutations. The fast detection of single base substitutions in small amounts of DNA has great potential in pretreated diagnosis and in oncology.

We thank Dr. Matthew Meyerson for critical reading of the manuscript and Naoya Hosono and Atsuko Miyazaki for their excellent technical assistance.