All lung cancer cell lines used in the present study originated from adenocarcinoma. The 11–18 cell line was obtained from the Cell Resource Center for Biomedical Research (Tohoku University, Sendai, Japan). The Ma1 cell line was provided by Dr Hirashima (Osaka Prefectural Habikino Hospital, Osaka, Japan). The A549 cell line was purchased from the American Type Culture Collection (Rockville, MD, USA). The EGFR mutation status of these cell lines was examined in our previous study (22). Cells were maintained in DMEM (Wako, Osaka, Japan) supplemented with 10% fetal bovine serum (Life Technologies, Carlsbad, CA, USA), 50 U/ml penicillin, and 50 U/ml streptomycin (both from Wako). Genomic DNA was prepared using a Wizard^® Genomic DNA Purification kit (A1120; Promega, Madison, WI, USA) according to the manufacturer’s instructions.

Ethical approval was obtained from the Tottori University Hospital and fully informed written consent was obtained from all patients involved prior to the surgery or tissue collection.

Tumor tissues were obtained from surgical specimens of resected tumors, from 143 lung cancer patients treated at Tottori University Hospital; these samples were embedded in Tissue-Tek OCT Compound (Sakura Finetechnical, Tokyo, Japan), and were immediately frozen at −80°C. Macrodissection of the OCT-embedded tissue samples was performed to enrich the final proportion of the tumor DNA, and DNA was extracted using the Wizard^® Genomic DNA Purification kit. For samples with discordant results between direct sequencing and mutation-specific PCR, a PCR-Invader method was performed by BML, Inc. (Tokyo, Japan) as a reference test.

For direct sequence analysis exon 19 and 21 of EGFR, the following PCR primers were used: EGFR exon 19F, 5′-GCAATATCAGCCTTAGGTGCGGCTC-3′ and EGFR exon 19R, 5′-CATAGAAAGTGAACATTTAGGAT GTG-3′; and EGFR exon 21F, 5′-CTAACGTTCGCCAGCC ATAAGTCC-3′ and EGFR exon 21R, 5′-GCTGCGAGCTCA CCCAGAATGTCTGG-3′. The PCR conditions were as follows: 1 cycle at 94°C for 9 min, followed by 40 cycles each consisting of 94°C for 1 min, 57°C for 1 min, and 72°C for 2 min, and a final cycle at 72°C for 5 min. The PCR products were purified with a MultiScreen-PCR filter plate (Millipore, Tokyo, Japan) and then sequenced using a BigDye Terminator v3.1 cycle sequencing kit and an ABI PRISM 3130xl genetic analyzer (Applied Biosystems, Foster City, CA, USA).

We designed a deletion-specific primer for the delE746-A750 mutation within exon 19 and a point mutation-specific primer for the L858R mutation within exon 21 of EGFR. Sequences of the primer sets were as follows: PCR forward primer for delE746-A750, 5′-CACAATTGCCAGTTAACGTCTTC-3′ (19DF) and PCR reverse primer for delE746-A750, 5′-TGTTGGCTTTCGGAG ATGTTTTG-3′ (19DR3); PCR forward primer for L858R, 5′-TCCCATGATGATCTGTCCCT-3′ (21F2f) and PCR reverse primer for L858R, 5′-CACCCAGCAGTTTGGTCC-3′ (21ARMS3).

For conventional PCR amplification using the mutation-specific primer sets, PCR conditions were as follows: the reaction mixtures contained 2 μl of 10X PCR buffer, 0.5 μl of dNTPs, 1 μl of each allelic-specific primer (10 μM), 0.2 μl of AmpliTaq^® Gold DNA polymerase (Applied Biosystems), 1 μl of template DNA, and 14.3 μl of ddH₂O in a total volume of 20 μl. Thermal cycling conditions on a PCR Thermal Cycler Dice (Takara, Shiga, Japan) were as follows for the delE746-A750 mutation: 1 cycle at 94°C for 9 min, followed by 35 cycles each consisting of 94°C for 1 min, 59°C for 1 min, and 72°C for 2 min, and a final cycle at 72°C for 5 min. Similarly, for the L858R mutation, conditions involved 1 cycle at 94°C for 9 min, followed by 35 cycles at 94°C for 1 min, 64°C for 1 min, and 72°C for 2 min, and a final cycle at 72°C for 5 min. The PCR products were then electrophoresed on agarose gels and stained with ethidium bromide.

For ultrarapid PCR-based EGFR mutation detection, we utilized a newly developed high-speed real-time PCR machine, termed the ‘Hyper-PCR’ UR104MK III (Trust Medical, Hyogo, Japan), which was jointly developed by ourselves and Trust Medical (23). The UR104MK III employs a novel temperature control technology. In this system, the PCR mixture is enclosed in a small vessel on a thin, flexible plastic disk and sealed with adhesive film, and the disk is rotated rapidly onto three separated heat elements. By controlling the speed of rotation and the temperature of the three heat elements, rapid PCR can be accomplished. Real-time monitoring of the fluorescent dsDNA dye produced during PCR progression, and the ability to perform melting curve analysis of the PCR product are also incorporated into this machine (Fig. 1). The typical time for amplification and detection when using this apparatus is <10 min.

The optimized reaction mixtures for use with this machine contained 1.6 μl of 10X Fast Buffer I, 1.3 μl of a 2.5 mM dNTP mixture, 0.4 μl of each allele-specific primer (10 μM), 0.2 μl of SpeedSTAR HS DNA Polymerase (5 U/μl) (Takara), 1 μl of template DNA, 1.6 μl of 1:2,000 SYBR^®-Green I nucleic acid gel stain (Cambrex Biosciences, Rockland, ME, USA), and 9.5 μl of ddH₂O in a total volume of 16 μl. Furthermore, dimethyl sulfoxide (DMSO) Hybri-Max^® (Sigma, St. Louis, MO, USA) was added to a final concentration of 5%. Thermal cycling conditions for ultrarapid PCR were as follows for the delE746-A750 mutation: 1 cycle at 94°C for 1 min, followed by 35 cycles each including 98°C for 1.30 sec, 55°C for 5.00 sec, and 72°C for 3.00 sec. Similarly, for the L858R mutation, conditions entailed 1 cycle at 94°C for 1 min, followed by 30 cycles each consisting of 98°C for 1.30 sec, 68°C for 8.00 sec, and a further 68°C for 8.00 sec. Total PCR cycling time for the delE746-A750 and the L858R mutation detection was within 6 and 9 min, respectively. Following PCR cycling, melting curve analysis was performed within 4 min.

For interpretation of the ultrarapid PCR results, criteria used in other studies of qualitative real-time PCR analysis were applied (24). In brief, to be considered as a positive result, a fluorescence signal generated during ultrarapid PCR should display an exponential amplification above the threshold level and the obvious crossing point (Cp) (25), with a single peak upon melting curve analysis, giving a unique melting temperature (Tm) value. A signal was considered as negative when no Cp value was obtained within the amplification cycles.

We first established a specific PCR to detect mutations within exons 19 and 21 of EGFR, which are representative mutations underlying the responsiveness of NSCLC to EGFR inhibitors (3). The genomic sequence of EGFR was retrieved from the NCBI database (NM_005228). For the in-frame deletion within exon 19, which removes nucleotides 2235–2249, causing a deletion of amino acids 746 through 750 (delE746-A750), we designed a deletion-specific primer, 19DR3 (Fig. 2A). This primer was designed to anneal only to the genomic sequence harboring the nucleotide 2235–2249 deletion, by connecting the flanking sequences on either side of the deletion. By shortening the 3′-end of the primer that corresponded to the upstream genomic sequence, we could improve the specificity of the primer.

For the exon 21 amino acid substitution, in which G is substituted for T at nucleotide 2573, causing an amino acid substitution of L to R (L858R), a point mutation-specific primer, 21ARMS3, was designed (Fig. 2B). Here, we employed the ARMS technique, and designed the primers to be refractory to PCR amplification of non-matching target sequences (26,27), by also including an additional mismatch in the candidate point mutation-specific primers (ARMS1-10) at positions −2 or −3 from the 3′-end of the primers (data not shown). Among these candidate primers, we chose the primer (ARMS3) that allowed discrimination without decreasing the sensitivity and specificity of amplification in a series of experiments in which these candidate primers were tested under the same temperature and time conditions in ultrarapid PCR.

The forward primers for each mutation-specific primer were designed to match the stable area of each EGFR (19F and 21F). The concentration of the PCR primers and magnesium, the annealing temperature and other cycling parameters, and the type of DNA polymerase used were determined by exploration, and the conditions described here are the final optimized conditions.

To evaluate the specificity of the mutation-specific primers, ultrarapid PCR using the mutation-specific primers was performed on lung cancer cell lines, and the concordance of these results with those of conventional PCR was evaluated. EGFR genotyping of the lung cancer cell lines Ma1, 11–18 and A549 had been performed by sequence analysis in our previous study, and were revealed as delE746-A750, L858R and wild-type, respectively (22). As shown in Fig. 3A, using primer sets, 19DF and 19DR3, a significant increase in fluorescence intensity was observed upon ultrarapid PCR only in Ma1 cells that harbor delE746-A750, and melting curve analysis revealed a clear peak at the expected Tm (81.3°C). The PCR product was visualized by agarose gel electrophoresis and the expected product of 113 bp was detected. To validate the data, conventional PCR was performed using the same primer sets, and a product of the same size was detected only in Ma1 cells (Fig. 3B).

A similar experiment was performed using the L858R point mutation-specific primers (21F2f and 21ARMS3) in ultrarapid PCR. As shown in Fig. 3C, a significant increase in fluorescence intensity and a clear melting curve peak at the expected Tm (87.1°C) was observed only in 11–18 cells that harbor the exon 21 L858R point mutation, and the size of the amplicon was confirmed as 166 bp by agarose gel electrophoresis. The same result was obtained by conventional PCR using this primer set (Fig. 3D). These data revealed that the combination of mutation-specific primers and ultrarapid PCR yielded satisfactory discrimination of the mutant alleles.

To evaluate the sensitivity of the assay, a serial dilution of lung cancer cell lines carrying EGFR mutations (Ma1 or 11–18) into wild-type cell lines (A549) was analyzed by ultrarapid PCR using the mutation-specific primers. Using the delE746-A750 primers in ultrarapid PCR, a mixture containing 0.1% Ma1 cells could be detected as positive from the amplification curve and melting curve analysis, and this result was confirmed by gel electrophoresis (Fig. 4A). The same result was also obtained by conventional PCR (Fig. 4B). Using the L858R primer sets, a cell mixture containing 1% 11–18 cells was detected as positive by ultrarapid PCR, and the result was confirmed by gel electrophoresis of the PCR product (Fig. 4C). The same detection limit was also observed with this primer set with conventional PCR (Fig. 4D). These data revealed that ultrarapid PCR combined with delE746-A750 and L858R mutation-specific primers allowed detection of 0.1 or 1% mutation-carrying lung cancer cells among wild-type cells, respectively.

We evaluated the concordance of the results obtained by direct sequencing and using mutation-specific primers in ultrarapid PCR or conventional PCR in 143 lung cancer tumors. Overall, the results of 138 of 143 samples were concordant among direct sequencing, ultrarapid and conventional PCR (Fig. 5). The remaining five samples that demonstrated inconsistent results are shown in Table I. In these samples, direct sequencing could not detect any mutations, whereas the results with ultrarapid PCR and conventional PCR were concordant: two delE746-A750 and three L858R mutations were identified. To evaluate these discrepant samples, we used PCR-Invader analysis, which confirmed the PCR-based results. Therefore, we concluded that the indicated mutations were indeed present in these samples, yet could not be detected by direct sequencing, possibly due to the lower sensitivity of detection of direct sequencing methodology.

When compared with the concluded EGFR mutation status, the sensitivity and specificity of direct sequencing was 84.8 and 100% respectively, whereas those of ultrarapid PCR were 100 and 100%, respectively (Table II). These data revealed that ultrarapid PCR has superior sensitivity and specificity in clinical samples to direct sequencing, and parallels that of the PCR-Invader method, which is one of the commonly used PCR-based methodologies in current clinical practice.

A 39-year-old woman with no history of smoking was referred to our hospital due to the presence of an abnormal shadow in her left upper lung field that was noticed during her regular medical checkup (Fig. 6A, upper panel). A computed tomography (CT) scan of the chest of this patient revealed a spiculated nodular shadow with a 35-mm diameter in the superior division of the left upper lobe (Fig. 6A, lower panel). F-2-deoxy-2-fluoro-D-glucose (FDG)-positron emission tomography revealed multiple lesions with high uptake in her liver and vertebrae. Based on the suspicion of adenocarcinoma with multiple metastases, we preformed flexible bronchoscopic examination with washing, brushing and forceps biopsy. In addition to the cytological and histological examination of the lavage and biopsy specimens, we applied ultrarapid PCR analysis to the lavage sample. The lavage was centrifuged and DNA was extracted within 40 min after sample collection. Ultrarapid PCR analysis subsequently revealed the presence of the delE746-A750 EGFR mutation in her lavage sample, within 6 min (Fig. 6B). After waiting for confirmation of positive results by cytology and histology, which were obtained 3–4 days after bronchoscopic examination, the patient was diagnosed as having adenocarcinoma with an EGFR mutation (cT2aN0M1b, stage IV). She commenced treatment with 150 mg erlotinib/day immediately, and her lung CT scan at 6 weeks after the initiation of treatment revealed a marked improvement.