Biotinylated Probe Isolation of Targeted Gene Region Improves Detection of T790M Epidermal Growth Factor Receptor Mutation via Peptide Nucleic Acid–Enriched Real-Time PCR

Abstract

BACKGROUND

The presence of the EGFR (epidermal growth factor receptor) T790M mutation in tumor tissue or body fluids from patients treated with EGFR tyrosine kinase inhibitors may indicate the onset of resistance to treatment. It is important to identify this mutation as early as possible so that treatment can be modified accordingly or potential side effects of further treatment can be avoided. This requirement calls for high detection sensitivity. Peptide nucleic acids (PNAs) are used as PCR clamps to inhibit amplification of wild-type DNA during PCR cycling, thereby enriching for rare mutations such as T790M. We describe a modification that improves the detection limit of PNA-clamp methods by at least 20-fold.

METHODS

We enriched the target by exposing genomic DNA to an EGFR exon 20–specific biotinylated oligo-nucleotide, followed by binding to streptavidin beads. We then prepared serial dilutions of the isolated target DNA containing the T790M mutation by mixing with wild-type DNA and then performed PNA clamp–based, real-time TaqMan PCR. For comparison, we performed PNA clamp–based PCR directly on genomic DNA.

RESULTS

Whereas the detection limit for PNA clamp–based PCR performed directly on genomic DNA is 1 mutant allele in 1000 wild-type alleles, conducting the assay with biotinylated oligonucleotide–enriched target DNA improved the detection limit to 1 mutant allele in 40 000 wild-type alleles. A possible explanation for the improvement in detection is that biotin-based target isolation efficiently eliminates wild-type DNA; therefore, fewer erroneous amplifications of wild-type DNA can occur early during the PCR.

CONCLUSIONS

Combining target molecule isolation via a biotinylated probe with PNA-enriched TaqMan real-time PCR provides a major improvement for detecting the EGFR T790M resistance mutation.

Non–small-cell lung cancer (NSCLC)³ patients harboring EGFR⁴ (epidermal growth factor receptor)-activating mutations are initially responsive to small-molecule tyrosine kinase inhibitors (TKIs) such as gefitinib and erlotinib; however, almost all of these patients develop drug resistance after prolonged treatment (1). For example, T790M, an acquired mutation that renders NSCLC patients resistant to gefitinib or erlotinib, is found in approximately 50% of tumors from patients who have acquired resistance to these kinase inhibitors (1–4). The presence of the T790M mutation in a primary lung tumor is considered clinically important and is a contraindication to the use of small-molecule TKIs for treatment (1–4). Similarly, the emergence of the T790M mutation in plasma from patients treated with TKIs is an indication of developing resistance (5).

Methods have been developed to detect the T790M mutation, with detection limits ranging from 1 mutant allele in 10 wild-type alleles to 1 in 2000 (6–9). These methods include a peptide nucleic acid (PNA)-clamping method (10, 11). PNAs have been widely used in PCR reactions as a “clamp” to inhibit the amplification of wild-type DNA and thereby enrich for mutant alleles (12). Detection limits for this method currently range from 1 mutant allele in 100 wild-type alleles to 1 in 1000 (10, 11). There is an increasing need, however, to further improve the detection of EGFR T790M because DNA with the T790M mutation can be initially present as a minor clone in a cancer that subsequently becomes drug resistant (13). It would be important to be able to screen tumors from NSCLC patients for the existence of the T790M mutation before they undergo TKI treatment, because patients with evidence of T790M may derive an even briefer benefit from EGFR TKI therapy (14). Furthermore, the detection of EGFR T790M in a lung cancer before any therapy could change the treatment from gefitinib to an irreversible EGFR inhibitor that is effective against the T790M mutation (15).

We demonstrate that biotinylated probe–based purification of a target from genomic DNA produces a major improvement in PNA-PCR–based mutation detection by increasing the sensitivity for detecting the mutation to at least 1 mutant allele in 40 000 wild-type alleles.

Reference human male genomic DNA was purchased from Promega and used as the wild-type DNA in dilution experiments with DNA containing the T790M mutation (H1975 cell line). DNA was extracted from cell lines with the QIAamp DNA Blood Maxi Kit (Qiagen). Primers and probes were synthesized by Integrated DNA Technologies. DNA was quantified with a NanoDrop spectrophotometer (Thermo Scientific).

Real-time PCR reactions were performed in a 25-μL volume in the presence of a dye that binds to the DNA minor grove (LCGreen Plus⁺; Idaho Technology) and with 20 ng genomic DNA prepared directly from cell lines. To quantify copies of EGFR exon 20, we serially diluted genomic DNA with human male genomic DNA calibrator. The final concentrations of the other PCR reagents were as follows: 1× GoTaq Flexi Buffer (Promega), 0.63 U of GoTaq Flexi DNA polymerase (Promega), 0.2 mmol/L of each de-oxynucleoside triphosphate, 0.2 μmol/L forward primer (5′-GCTGGGCATCTGCCTCA-3′), 0.2 μmol/L reverse primer (5′-CAGGAGGCAGCCGAAGG-3′), 2.5 mmol/L MgCl₂, and 0.1× LC-Green Plus⁺. The size of the PCR amplicon is 67 bp. The PCR cycling was performed on a Cepheid SmartCycler^™ machine as follows: 95 °C for 120 s and 50 cycles of 95 °C for 15 s and 60 °C for 30 s (fluorescence reading on), followed by DNA melting from 60 °C to 95 °C at a temperature-ramping rate of 0.2 °C/s. Genomic DNA containing at least 10⁶ copies of EGFR exon 20 molecules were digested in a 480-μL reaction volume containing 1× NEBuffer 2 (New England Biolabs), 1× BSA, 100 U each of RsaI, EcoRI, BamHI, and EcoRV. The reaction was incubated at 37 °C for 3 h and then purified in an Ultracel YM-30 Microcon column (Millipore). DNA was eluted in 30 μL water. To enrich for EGFR exon 20, we added a 5′-biotinylated probe (biotin–Spacer 18–GCCTGCTGGGCATCTGCCTCACCTCCACCG; synthesized by Integrated DNA Technologies) to the purified digested genomic DNA and diluted the probe to a final concentration of 33 nmol/L with 6× SSPE Buffer (American Bioanalytical) in a 45-μL volume. The mixture was denatured at 100 °C for 2 min in a PCR thermocycler and then quickly cooled on ice for 5 min. The hybridization was then performed in a thermocycler for 16 h at 58 °C. The hybridization mixture was then purified with an Ultracel YM-30 Microcon column and eluted in 40 μL water. We washed 10 μL Dynabeads M-270 Streptavidin (Invitrogen) 3 times with 1× binding and washing buffer (5 mmol/L Tris-HCL, pH 7.5, 0.5 mmol/L EDTA, and 1 mol/L NaCl) and resuspended the beads in 40 μL of 2× binding and washing buffer. The 40-μL hybridization mixture of purified probe and target was captured by mixing it with the 40 μL of processed Dynabeads and incubating the mixture on a shaker for 1 h at room temperature. The beads were washed 3 times with 1× binding and washing buffer supplemented with 1 mL/L Tween 20, twice with 1× binding and washing buffer, and once with water. Finally, the beads were resuspended in 15 μL water, denatured at 95 °C for 2 min, and placed immediately on DynaMag magnets (Invitrogen). The suspension was recovered for further analysis.

The PNA-clamp reaction components were similar to those for the LCGreen dye quantification of EGFR exon 20 described above, except for the use of a locked nucleic acid–modified TaqMan probe and PNA clamp, as previously reported (11), instead of the LCGreen dye. The final concentrations of the TaqMan probe and PNA were both 0.2 μmol/L. Approximately 100 ng of genomic DNA or, alternatively, 2 μL purified target were used for T790M detection. The PCR cycling conditions were 95 °C for 120 s and 70 cycles of 95 °C for 15 s and 60 °C for 30 s (fluorescence reading on). PCR products were purified with exonuclease I (New England Biolabs) and shrimp alkaline phosphatase (USB/Affymetrix). We then used primer 5′-40T-GCTGGGCATCTGCCTCA-3′ (Integrated DNA technologies) to sequence the purified product by the Sanger method. Precautions were taken to minimize carryover contamination, and PCR tubes were prepared and uncapped in a negative-flow hood. All experiments were repeated from starting material at least 4 independent times.

As shown in Fig. 1A in the Data Supplement that accompanies the online version of this Brief Communication at http://www.clinchem.org/content/vol57/issue5, genomic DNA is first digested by a restriction endonuclease (RsaI in this case) into fragments, one of which contains the entire target region between 2 closely spaced restriction sites. The duplexes containing the hybridized probe and EGFR exon 20 are bound to the streptavidin-coated magnetic beads and washed. EGFR exon 20 single-stranded DNA is released from the beads by brief heat denaturation. A PNA clamp–based real-time PCR with a TaqMan probe containing a locked nucleic acid matching the T790M mutation [as described by Miyazawa et al. (11)] is then used to quantify the T790M mutation (see Fig. 1B in the online Data Supplement). With the biotinylated probe, we achieved an enrichment in EGFR exon 20 of at least 1000-fold relative to other genomic DNA fragments. This enrichment estimate is based on the amount of isolated product and results obtained with input genomic DNA in a TaqMan-based real-time PCR targeting the GAPDH (glyceraldehyde-3-phosphate dehydrogenase) gene (data not shown). This result translates into a recovery efficiency for EGFR exon 20 of approximately 20%–30% relative to the input genomic DNA.

We diluted genomic DNA from the H1975 cell line with the T790M mutation by 1000- to 40 000-fold with wild-type DNA. To enable the detection of low-level mutations, we used genomic DNA containing approximately 10⁶ copies of EGFR exon 20 for isolation via the biotin probe. To evaluate the reproducibility of the biotin-isolation process and to enable a fair comparison of the mutation-detection limit of PNA clamp–based PCR, we measured the copy number of EGFR exon 20 several times in independent experiments before performing the PNA-PCR amplification. LCGreen-based real-time PCR analyses of EGFR exon 20 fragments containing the T790M mutation isolated from dilutions into wild-type DNA (1000- to 40 000-fold dilution) demonstrated equal numbers of copies of EGFR exon 20 (see Fig. 2 in the online Data Supplement). The LCGreen dye does not discriminate between wild-type and mutant sequences and therefore can be used to quantify EGFR exon 20 copy number for all samples. The real-time PCR calibration curve was established with a serially diluted human control genomic DNA of known concentration, as we have previously described (9).

We used equal numbers of EGFR exon 20 copies isolated from wild-type DNA to test serially diluted T790M mutations via PNA probe–enriched real-time PCR. When genomic DNA is used as the starting material (Fig. 1A), the detection limit is 1 T790M allele in 1000 wild-type alleles, a result in agreement with Miyazawa et al. (11). The inability to distinguish lower concentrations of the mutation is due to increased background signals (false positives) generated by the wild-type DNA (Fig. 1A). In contrast, the approach involving isolating the biotin target can clearly detect T790M mutations at concentrations as low as 1 mutant allele in 40 000 wild-type alleles (Fig. 1B), because the wild-type samples are reproducibly not generating false-positive signals.

Fig. 1. Comparison of PNA-enriched TaqMan real-time PCR with biotin-isolated EGFR exon 20 molecules and the same technique with whole genomic DNA as starting material.

(A), Quantification of the EGFR exon 20 T790M mutation by PNA-enriched TaqMan real-time PCR directly from genomic DNA. (B), Quantification of the EGFR exon 20 T790M mutation by PNA-enriched TaqMan real-time PCR from a biotin-purified target. 1:1000 indicates 1 T790M allele in 1000 wild-type alleles. NTC, no-template control.

To better understand the origin of the background signals from wild-type samples when starting from genomic DNA, we used genomic DNA and biotin-isolated starting DNA and sequenced the PNA-PCR product from the 1:1000 (i.e., 1 mutant allele among 1000 wild-type alleles) dilution and wild-type samples. Wild-type DNA from genomic DNA produces false-positive T790M signals (see Fig. 3 in the online Data Supplement), possibly because of polymerase misincorporations in early PCR cycles; however, wild-type DNA from biotin-isolated DNA does not show the T790M C>T mutation at this position. Using PNA-PCR may require several PCR cycles to accumulate enough copies of the target molecules for efficient binding by the PNA probe, which may allow polymerase misincorporations to occur in wild-type DNA samples when starting from genomic DNA, but not when biotin probes are used.

Target purification via biotinylated probes has previously been used to increase the number of starting target copies before constant denaturant capillary electrophoresis or restriction fragment length polymorphism analysis (16, 17). The present work demonstrates that for methods such as PNA-PCR that operate by enriching the mutation during PCR amplification, biotin probe isolation of the target further improves PCR-based mutation detection by avoiding PCR errors at early PCR cycles. Because of the relatively large amount of starting genomic DNA required for the present procedure (approximately 5 μg), this approach is intended primarily for screening tumor tissue or cell lines for rare mutations. The present improvement should also be applicable for detecting other PNA clamp–enriched mutations or to alternative methods that enrich mutations during the PCR, such as restriction endonuclease–mediated selective PCR (18, 19) or coamplification at lower denaturation temperature–PCR (20, 21).