Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2015 Apr 1.
Published in final edited form as: Clin Chem. 2014 Oct 8;61(1):267–277. doi: 10.1373/clinchem.2014.228361

SINGLE-TUBE, HIGHLY PARALLEL MUTATION ENRICHMENT IN CANCER GENE PANELS USING TEMPERATURE-TOLERANT-COLD-PCR

E Castellanos-Rizaldos 1, K Richardson 2, R Lin 2, G Wu 2, G M Makrigiorgos 1,3,*
PMCID: PMC4281501  NIHMSID: NIHMS640207  PMID: 25297854

Abstract

BACKGROUND

Multiplexed detection of low-level mutations presents a technical challenge for many technologies, including cancer gene panels used for targeted-re-sequencing. Analysis of mutations below ~2–5% abundance in tumors with heterogeneity, samples with stromal contamination, or biofluids, is problematic due to increased ‘noise’ from sequencing errors. Technologies that reduce noise via deep-sequencing unavoidably reduce throughput and increase cost. Here we provide proof-of-principle that COLD-PCR technology enables multiplex low-level mutation detection in cancer gene panels while retaining throughput.

METHODS

We have developed a multiplex temperature-tolerant-COLD-PCR (fast-TT-COLD-PCR) approach that uses cancer gene panels developed for massively parallel sequencing. Following a multiplex pre-amplification from genomic DNA we attach tails to all amplicons and perform fast-TT-COLD-PCR. This approach gradually increases denaturation temperatures in a step-wise fashion, such that all possible denaturation temperatures are encompassed. By introducing modified nucleotides, fast-COLD-PCR is adapted to enrich for Tm-increasing as well as Tm-decreasing mutations over all amplicons, in a single tube.

RESULTS

Using custom-made and commercial gene panels containing 8, 50, 190 or 16,000 amplicons we demonstrate that fast-TT-COLD-PCR enriches mutations on all examined targets simultaneously. Incorporation of dITP/dDTP in place of dGTP/dATP enables enrichment of Tm-increasing mutations. Serial dilution experiments demonstrate a limit-of-detection of ~ 0.01–0.1% mutation abundance using Ion-Torrent and 0.1–0.3% using Sanger sequencing.

CONCLUSIONS

Fast-TT-COLD-PCR improves the limit of detection of cancer gene panels by enabling mutation enrichment in multiplex, single tube reactions. This novel adaptation of COLD-PCR converts subclonal mutations to clonal, thereby facilitating detection and subsequent mutation sequencing.

Keywords: multiplex, COLD-PCR, Temperature-tolerant-COLD-PCR, mutation scanning, mutation enrichment

INTRODUCTION

A low abundance of cancer-associated somatic mutations can occur in multiple situations such as in heterogeneous tumor samples with variant subclones, in surgical tumor specimens containing stromal tissue, in tumor derived DNA found in biofluids, and in samples taken from tumor margins. The ability to monitor this subclonal population may have a profound impact in understanding therapy outcome, and cancer remission or relapse (1–3). Targeted amplicon re-sequencing has emerged as a practical opportunity for integrating technology and clinical oncology (4–8). However the detection of low-prevalence mutations using targeted re-sequencing presents formidable technical challenges (9) and is fraught with sequencing errors leading to false-positive DNA changes that preclude reliable mutation calling, irrespective of coverage. This sequencing ‘noise’ is often independent of interrogation depth (10) hence rare mutations in plasma, circulating tumor cells (CTCs) and mixed clinical samples cannot be sequenced reliably. Current consensus is that the limit of detection (LOD) for variant identification in targeted re-sequencing panels is a ~ 5% mutant-to-wild type ratio, or mutation abundance (4–7). Recent enhancements enable NGS to reduce the noise substantially, via multiple short reads of few sequences (11, 12), random barcoding of DNA molecules on one (13, 14) or on both DNA strands (15) or via bio-informatics approaches (16–19). However, all these methods require very high coverage for correct calling of mutations in a limited number of target sequences. Increasing the depth of coverage is unavoidably accompanied by a decrease in NGS throughput since every sequence has to be read multiple times, consuming resources while yielding little information, which leads to an increase in cost (20).

Here we present an alternative approach that enables detection of very low level mutations in cancer gene panels without requiring excessive depth of coverage. In this approach, mutations are first magnified via CO-amplification at Lower Denaturation temperature PCR (21) prior to sequencing. A modification of COLD-PCR, temperature-tolerant-fast-COLD-PCR (fast-TT-COLD-PCR (22)) that enables mutation enrichment in amplicons having distinct melting temperatures using a single PCR program, was employed. Here we describe a new, multiplexed fast-TT-COLD-PCR adaptation that enables mutation enrichment of numerous amplicons within a single-tube. Multiplexed fast-TT-COLD-PCR operates by attaching common primers for amplification to all sequences. Addition of modified nucleotides during fast-COLD-PCR expands the types of mutations that can be enriched. We demonstrate that fast-TT-COLD-PCR allows reliable sequencing down to ~0.01–0.3% mutation abundance, compared to a lower limit of detection of ~2–5% using conventional PCR.

MATERIALS AND METHODS

DNA from clinical samples and cell lines

Human genomic DNA (Promega, Inc., Madison, WI, USA) was mixed with DNA from multiple mutated cell lines (Supplemental Table 1), to generate DNA with mutations at an abundance of ~5% at several sequence positions. Subsequent dilution with wild-type DNA was used to create lower mutation abundances. Genomic DNA was isolated using the DNAeasy™ Blood & Tissue Kit (Qiagen), following manufacturer’s instructions. Blood from healthy individuals was obtained in accordance with the Internal Review Board of Dana-Farber Cancer Institute. Cell-free circulating DNA (cfDNA) was isolated from plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen).

Optimization of fast-COLD-PCR at a single critical denaturation temperature (Tc)

We used fast-COLD-PCR (21, 23–25) at a single Tc (Tc = Tm −1), as previously observed (26) to evaluate the levels of enrichment of eight mutant sequences, and then tested the same targets by using a multiplex fast-TT-COLD-PCR approach. Eight genes were evaluated (Supplemental Table 2 and Supplemental Table 3). Tm values were predicted in silico using uMelt (27) and verified experimentally (Supplemental Table 4).

DNA pre-amplification using custom or commercial primer pools, prior to fast-TT-COLD-PCR

Mutation enrichment using custom primer pools was evaluated in both 8-plex and 50-plex reactions. Amplicon-specific primers contained two common sequences, Tag1 and M13 at their 5’ends, in order to generate amplicons with uniform ends. For pre-amplification the KAPA Hifi DNA polymerase system was used, as previously reported (28) (error rate of 2.8×10−7, KAPA Biosystems, USA). Following pre-amplification, fast-TT-COLD-PCR reactions were performed using Tag1 and M13 primers. Cycling conditions, target genes and primer sequences are summarized in Supplemental Tables 2, Supplemental Table 3 and Supplemental Table 5. Amplified products were treated with exonuclease I (New England Biolabs) and shrimp alkaline phosphatase (Affymetrix) and sequenced at Eton Bioscience (Cambridge, USA). Chromatograms were analyzed using BioEdit v7.1.3 (Ibis Biosciences). Number of mutant alleles relative to wild type alleles were estimated using peak height values of the Sanger sequencing chromatograms, as reported (29). Mutation enrichment at all eight examined targets was assessed via Sanger sequencing chromatogram comparison.

Genomic DNA containing ~5% mutated alleles was also pre-amplified using Ion Ampliseq™ with either the cancer primer pool (comprising 46 genes and 190 amplicons, interrogating 739 somatic mutations) or the comprehensive cancer panel (>400 oncogenes, ~ 16,000 amplicons with all-exon coverage of 409 genes), according to the Ion Ampliseq™ library preparation user guide (Life Technologies, USA). In-house-prepared ssDNA adaptors 12-mers and 24-mers were added to the phosphorylated amplicons, and ligation was applied using the dsDNA adaptors as described (30). Fast-TT-COLD-PCR was then applied using the protocols described in Supplemental Table 2 and Supplemental Table 4. Either one or two successive rounds of Fast-TT-COLD-PCR were applied to increase the mutation enrichment (31, 32). Mutation enrichment at the examined targets was assessed via Sanger sequencing chromatogram comparison, and via high resolution melting (HRM) analysis on a Lightscanner HR96 system (Idaho technologies Inc). Experiments were repeated in triplicate.

Ion Torrent sequencing of multiplex fast-TT-COLD-PCR products

Following fast-TT-COLD-PCR, mutation-enriched amplicons from each temperature window were processed using the standard barcoding and library preparation protocol for Ion Torrent sequencing. Libraries with ligated Ion Torrent adapters were assessed for DNA quality and quantity on Agilent bio-analyzer, and then pooled together into a single tube prior to Ion Torrent sequencing. An emulsion PCR containing Ion Sphere Particles (ISPs) and libraries with ligated Ion Torrent adapters were performed to optimally amplify amplicons bound to ISPss using Ion Torrent’s One Touch2™ instrument and Ion PGM Template OT2 200 Kit.

Following enrichment of template-positive Ion Sphere Particles, templates were sequenced with the Ion Torrent Machine (PGM using the 318 chip) per manufacturer’s protocol. Data analysis including base calling, read filtering, and demultiplexing were performed using the Torrent Suite software (version 3.4.2 or higher) from Ion Torrent. Variants were detected using the variant Caller plugin from Torrent Suite. Reads for each non-reference bases at the nucletotide of interest as well as the adjacent nucleotides were recorded and plotted in a noise plot to distinguish the true mutation versus background noise. Unfiltered NGS data (BAM files) were also loaded into Integrative Genome Viewer 2.3 (IGV, Broad Institute) (33) using human genome hg19 as reference.

Fast-COLD-PCR using modified nucleotides, deoxyinosine and 2,6-diaminopurine triphosphates

Reactions were carried out with the Biolase™ DNA polymerase system using 2.5 mM final concentration of MgCl2, 0.5 µM of primers, 5 U of polymerase and 0.1 mM of 2,6-diaminopurine triphosphate (dDTP, TriLink Biotechnologies, USA)/deoxyinosine triphosphate (dITP, Thermoscientific, USA)/ dCTP/dTTP, as reported by Suspene et al (34). Primer sequences and cycling conditions are summarized in Supplemental Table 2 and Supplemental Table 3.

Spiking of mutated amplicon into ligation-mediated PCR product from cfDNA, prior to fast-TT-COLD-PCR

cfDNA recovered from plasma was amplified by ligation mediated-PCR as described (30). A TP53 exon 8 gene region from DNA containing a 1% mutation (c.818G>A) was separately pre-amplified using primer set 115-P20 (Supplemental Table 3). Following quantification of the two samples, the mutated amplicon was spiked into the LM-PCR product such that the number of TP53 target copies resembled those present endogenously in cfDNA. A COLD-PCR reaction was then applied using the common primers P20. For the COLD-PCR reaction, we used 1U of Phusion® High Fidelity DNA polymerase, together with 5 µM of primer P20 (Supplemental Table 3).

RESULTS

Fast-TT-COLD-PCR principle

COLD-PCR in its simplest form (fast-COLD-PCR) implements a lower denaturation temperature (Tc, critical denaturation temperature) during an otherwise conventional PCR amplification, and thus allows preferential amplification of samples with mutations that decrease the amplicon Tm (21, 23–25, 31). Each amplicon has a distinct Tc that enables Tm-decreasing mutations anywhere on the amplicon to be enriched during fast-COLD-PCR. The constraint of using a distinct Tc per amplicon means it requires a different PCR protocol for each reaction and makes multiplexing during fast-COLD-PCR problematic. To address this issue, we recently developed TT-COLD-PCR (22). In this approach, denaturation temperatures are increased gradually in a step-wise fashion, which allows for a preferential enrichment of mutated sequences in diverse amplicons with different Tc using a single PCR protocol. Hence thermocyclers can process amplicons with distinct Tc on different wells using a single program. Here we develop this principle further to enable multiplexed fast-TT-COLD-PCR in a single tube.

In the single-tube approach (Figure 1) panels of amplicons (8–16,000 per reaction herein) are pre-amplified from genomic DNA using custom-made (28) or commercially available (Ion Ampliseq™) sets of primers. Next, common oligonucleotide 'tails' are attached to the ends of all amplicons via a blunt-end ligation reaction. Following ligation, the sample is split into eight separate tubes. A different fast-TT-COLD-PCR reaction then is applied to each tube using step-up denaturation temperature protocols (22) spanning different temperature windows. Each temperature step is 0.3°C and seven successive steps are employed covering a ~2°C temperature window. Mutations contained within amplicons whose Tc falls within the corresponding temperature window are enriched during fast-TT-COLD-PCR, while the remaining amplicons either do not amplify or do not enrich mutations during amplification. By employing a different temperature window for each of the eight reactions, all possible Tc values are encompassed, hence mutations in all amplicons are enriched in at least one of the reactions. The samples were first examined via Sanger sequencing. Alternatively, for examination via Ion Torrent sequencing, each of the eight PCR products was barcoded, following which samples were mixed and processed for library preparation. To examine mutation enrichment within each barcoded group of amplicons, we compared the detected mutation frequency to the mutation frequency in samples treated identically but amplified via conventional PCR.

Figure 1. Multiplex fast-Temperature-Tolerant-COLD-PCR using cancer amplicon panels.

Figure 1

Open in a new tab

Genomic DNA is pre-amplified using either custom or Ion Ampliseq™ cancer panels (190–16,000 target amplicons). After ligation of DNA adaptors, the sample is split in eight parallel reactions, and amplified via fast-TT-COLD-PCR using a single primer to co-amplify all targets. Each TT reaction is covering a distinct range of denaturation temperatures, such that eight reactions cover all denaturation temperatures in the pool of amplicons. Levels of mutation enrichment were assesed by Sanger Sequencing or by Ion Torrent sequencing.

Fast-COLD-PCR for single sequences at a single Tc

Eight single amplicons generated via conventional PCR from genomic DNA, containing ~5% – 10% mutation abundances were first evaluated for enrichment with fast-COLD-PCR at a single Tc, prior to using multiplex fast COLD-PCR, for comparison purposes. Tm of each of amplicon was predicted in silico, and then confirmed experimentally. Depending on the amplicon, mutation enrichment for single sequences ranged from 4 to 16-fold for the eight amplicons examined (Supplemental Figure 1).

8-plex and 50-plex fast-TT-COLD PCR using custom primer sets

Genomic DNA from cell lines with known mutations in several cancer relevant genes, including the eight targets examined above, were combined to form ~5% or 1% mutation abundances and was subjected to fast-TT-COLD-PCR, using an 8-plex or 50-plex reactions. Next, the pre-amplified sample was split into eight reactions and a separate fast-TT-COLD-PCR was applied in temperature windows shown in Supplemental Table 4 to encompass the full range of denaturation temperatures 84.3 to 94.5°C predicted for all amplicons present in the reactions. Two consecutive rounds of 8-plex- or 50-plex fast-TT-COLD-PCR led to mutation enrichment of 7-fold to 46-fold over all eight mutated target amplicons, as inferred by assessment of Sanger chromatograms (Figure 2, 50-plex reactions; and Supplemental Figure 2, 8-plex). The mutation enrichment obtained using single-tube mutltiplex fast-TT-COLD-PCR is comparable to the enrichment obtained for single amplicons (Supplemental Figure 1). This indicates that the presence of diverse amplicons in the same reaction does not influence substantially the mutation enrichment for any given sequence within the reaction. The levels of enrichment shown by Sanger Sequencing were also confirmed by Ion Torrent-based NGS (Supplemental Table 6). Generally good agreement was observed between the two sequencing technologies.

Figure 2. Sanger sequencing following 50-plex fast-TT-COLD-PCR or, alternatively, conventional 50-plex PCR.

Figure 2

Open in a new tab

A mixture of genomic DNA from several cell lines (~5% mutation abundance) was pre-amplified using a 50-plex PCR using gene specific primers containing common sequences on the 5’ end. The sample was then split in eight parallel tubes and fast-TT-COLD-PCR was then performed using a single primer approach to co-amplify all 50 amplicons.

Adaptation of fast-TT-COLD-PCR for enrichment of a Tm-increasing mutation

While Tm-lowering mutations (C:G>T:A, C:G>A:T) comprise the majority of somatic mutations in most major cancers (35), Tm-increasing mutations such as T:A>G:C or T:A>C:G are also prevalent in certain cancers including renal and ovarian (35). To allow fast-COLD-PCR to enrich Tm-increasing mutations we incorporated modified nucleotide triphosphates (dDTP and dITP) during fast-COLD-PCR. When incorporated into DNA during PCR, dITP pairs with cytosine via two hydrogen bonds and dDTP pairs with thymidine via three hydrogen bonds (34). The combination of hydrogen bonding and base stacking interactions result to a lower amplicon Tm whenever G is replaced by I and increasing the Tm when A is replaced by D. Application of these principles enabled us to reverse the effect of mutations on Tm, thereby enabling fast-COLD-PCR to be used for enrichment of Tm-increasing mutations in the same way it is used for Tm-decreasing mutations (Figure 3).

Figure 3. Selective enrichment for A:T>G:C or A:T>C:G somatic mutations by incorporating deoxyinosine and 2,6-diaminopurine triphosphates using fast-COLD-PCR.

Figure 3

Open in a new tab

(A) Principle of fast-COLD-PCR, where PCR occurs at a lower denaturation temperature. Sequences that contain mutations that decrease the melting temperature of the amplicon (G:C>T:A or G:C>A:T) amplify preferentially over wild type sequences. (B) Alteration of DNA hydrogen bonding rules by replacing dATP and dGTP by 2,6-diaminopurine triphosphate (dDTP) and deoxyinosine (dITP) during the PCR reaction. dDTP forms three hydrogen bonds with thymidine and dITP forms two hydrogen bonds with cytidine. (C) Enrichment of sequences that contain mutations that increase the melting temperature of the amplicon (T:A>G:C or T:A>C:G) by fast-COLD-PCR by using dDTP and dITP instead of dDTP and dITP.

An 87 bp amplicon containing a 5% mutation abundance with a Tm-increasing mutation (Supplemental Table 1; c.823T>G, TP53 exon 8 from PFSK-1 cells) was subjected to either fast-COLD-PCR at a single Tc, or to conventional PCR and followed by Sanger Sequencing. Mutation enrichment varied from ~4-to-7-fold, over a broad range of denaturation temperatures, while optimal Tc was observed at Tm-1.5, depicted on Figure 4, panel A. We then repeated evaluation of enrichning the same Tm-increasing mutation when amplifying a longer, 152 bp length amplicon. We observed that, under these conditions, mutation enrichment also ocurred over a broad range of temperatures, from Tm-2 to Tm. Optimal Tc was observed at Tm-1, depicted on Figure 4, panel B. We next demonstrated that this same strategy was possible to use for mutation enrichment, but in a four-plex fast -COLD-PCR or a 50-plex fast- COLD-PCR format (Supplemental Figure 3), and the outcome was comparable to that observed in Figure 4. To evaluate the efficiency of this approach in detecting lower levels of mutations, we amplified wild type DNA and mutation dilutions of 10%, 1%, 0.5%, 0.25% and 0.125%, with use of fast-TT-COLD-PCR in the presence of modified nucleotide triphosphates followed by HRM. All mutation dilutions showed distinct differential denaturation curves compared to the wild type sample, after two consecutive rounds of COLD-PCR, Figure 4, panel C.

Figure 4. Conventional or fast-COLD-PCR of a Tm-increasing mutation at a single Tc using modified nucleotides.

Figure 4

Open in a new tab

(A) DNA with a ~5% mutation abundance, T>G mutation, was amplified using either conventional PCR or fast-COLD-PCR reactions in the presence of dDTP and dITP. A ~7-fold enrichment of a Tm-increasing mutation is demonstrated. (B) Evaluation of enrichment over two rounds of conventional or fast-COLD-PCR for a longer amplicon (152 bp) in the presence of modified nucleotides. A ~10% mutation abundance, T>G mutation, was amplified using a conventional PCR or fast-COLD-PCR approach in the presence of dDTP and dITP. Optimal mutation enrichment was observed at Tc = Tm-1°C. (C). High Resolution Melting analysis using modified nucleotides (dITP, dDTP). We examined differential HRM curves of serial mutation dilutions amplified using fast-TT-COLD-PCR or conventional PCR against wild type DNA. Tm-increasing mutations down to ~0.12% abundance were distinguishable from wild type samples.

190-plex fast-TT-COLD-PCR using Ion Ampliseq™ primer sets

To employ fast-TT-COLD PCR in conjunction with commercial cancer panels used for targeted re-sequencing, we used the Ion Ampliseq™ cancer primer pool (190 amplicons) for pre-amplification from the same genomic DNA mix described in materials and methods. Five of our eight targets with known mutation positions fell within the amplicons generated by the commercial primers. Following multiplex PCR from genomic DNA and ligation of a common tail, we subjected the samples to fast-TT-COLD-PCR using the same approach applied for the custom-made 50-plex reactions. The level of mutation enrichment achieved with use of the 190 amplicon panels was evaluated by Sanger Sequencing in the target regions known to contain mutations. 190-plex fast-TT-COLD- PCR and conventional PCR were compared by examining a wild type sample and mutant serial dilutions (5%, 2.5%, 1.25%, 0.63%, 0.31% and 0.16%): a gene-specific PCR was performed after conventional PCR and fast-TT-COLD- PCR for exon 5 of TP53, followed by HRM. Conventional PCR followed by HRM had a limit of detection of 2.5% mutation abundance, consistent with previous reports (36). Fast-TT-COLD-PCR on the other hand discriminated differences from wild type DNA down to 0.31% mutation abundance (Supplemental Figure 4). Mutation enrichment assessed by Sanger sequencing for 190-plex fast-TT-COLD-PCR was 4-to-18-fold, depending on the amplicon and amplification strategy used (one or two rounds of fast-TT-COLD-PCR) (Figure 5).

Figure 5. Sanger sequencing after 190-plex fast-TT-COLD-PCR or, alternatively, conventional PCR.

Figure 5

Open in a new tab

A mixture of genomic DNA from several cell lines (~5% mutation abundance) was pre-amplified using Ion Ampliseq™ Cancer Primer pool (190 amplicons) and then processed with Ion Ampliseq library kit 2.0 incorporating modified conditions. Ion AmpliSeq™ adapters were replaced by custom linker sequences. The sample was split in eight parallel tubes and fast-TT-COLD-PCR was then performed using a single primer approach to co-amplify all 190 amplicons. Enrichment was evaluated after either one or two consecutive rounds of fast-TT-COLD-PCR versus conventional PCR.

Testing fast-TT-COLD-PCR with large amplicon pools: 16,000 amplicons and circulating DNA (genome-wide) pool

The data indicate that the performance of fast-TT-COLD-PCR with small to modest amplicon pools, 8–190 amplicons per reaction, stays relatively unaffected by the number of amplicons and the mutation enrichment remains approximately the same. To test whether even larger pools of amplicons can be used, we evaluated fast-TT-COLD-PCR with the commercial Ion Ampliseq™ comprehensive cancer panel primer pool (~16,000 amplicons). Following a similar testing approach as applied for the Ion Ampliseq™ cancer primer pool, we examined mutation enrichment following Sanger sequencing using fast-TT-COLD-PCR compared to using conventional PCR. Supplemental Figure 5 indicates that mutation enrichment of 3–8-fold were obtained on the target amplicons examined.

Finally we also evaluated the extent to which low level mutations could be enriched on a whole genome scale using fast-COLD-PCR and plasma-circulating DNA from a normal individual spiked with a mutation-containing DNA sequence (~1% mutation abundance was used). When fast-COLD-PCR is applied to the spiked sequence independently, without mixing with circulating DNA, an enrichment of ~58-fold is obtained (Panel A, Supplemental Figure 6). On the other hand when this sequence is spiked within circulating DNA and co-amplified a mutation enrichment of ~34-fold was obtained (Panel B, Supplemental Figure 6). Overall the data indicate that substantial mutation enrichment can be obtained using fast-COLD-PCR irrespective of the number of amplicons used in the reaction.

Mutation enrichment at low mutation abundances and limit of detection

Under COLD-PCR conditions, the lowest starting mutational abundances result to greatest mutation enrichment (36). Consistent with previous data, when samples containing ~1% mutational abundances were subjected to multiplexed fast-TT-COLD-PCR we observed increased fold enrichments of 46–56-fold (Supplemental Figures 2 and 6).

To ascertain the limit of detection when multiplexed fast-TT-COLD-PCR is followed by sequencing, we conducted a serial dilution study using 50-plex reactions. Figure 6 depicts representative results from two targets, TP53 exon 9 and EGFR exon 20, with mutational abundances of 0.012 – 1.25 % for which we tested the limit of detection using Sanger sequencing and NGS. Conventional PCR followed by either Sanger sequencing or NGS was not able to distinguish somatic variants present at these low-abundance levels. However, fast-TT-COLD-PCR followed by Sanger sequencing depicted limits of detection of ~0.1% (TP53 exon 9) and ~0.3% (EGFR exon 20). Using Ion Torrent based NGS in two independent runs we were able to call (p<0.001) somatic variants down to ~0.01% (TP53) and ~0.1% (EGFR) mutation abundance. At these mutation abundances the signal was still distinct from the noise levels (Figure 6).

Figure 6. Mutation enrichment at low mutation abundances and limit of detection (50-plex fast-TT-COLD-PCR).

Figure 6

Open in a new tab

(A) Sanger sequencing results of a serial dilution experiment for two target amplicons (TP53 exon 9 and EGFR exon 20). Conventional PCR was unable to detect mutations below ~10–20%, while fast-TT-COLD-PCR enabled mutation detection down to ~0.1–0.3% abundance. (B) Ion Torrent sequencing results from the same serial dilution experiment for TP53 exon 9 and EGFR exon 20. Conventional PCR was unable to detect mutations below ~5%, while fast-TT-COLD-PCR enabled mutation detection down to ~0.01–0.1% abundance.

DISCUSSION

Among recent developments providing methods to address the problem of sequencing noise (11–16) our group proposed the use of COLD-PCR to enrich mutations on single amplicons prior to NGS, enabling assessment of mutation frequencies down to 0.02–1% (10, 37). As originally conceived, this method had no multiplexing capability due to the fact that the pre-amplification step prior to NGS is performed under cycling conditions optimal for each individual target (10). The present development of single tube, multiplexed fast-TT-COLD-PCR removes this limitation and can potentially be incorporated within the workflow for targeted re-sequencing using cancer panels. As Figures 2, 5 and Supplemental Figure 5 indicate, the mutation enrichment obtained prior to sequencing when using fast-TT-COLD-PCR can be applied to large pools of amplicons, improving the limit of detection of sequencing technologies.

Application of fast-TT-COLD-PCR converts lower level mutations to modest or high-level mutations that should require fewer sequence reads to call lower level mutations. Indeed, per Flaherty et al (17) a 0.1% mutation abundance requires a read depth of at least 10,000 reads to reduce sampling error of the binomial distribution enough for reliable variant detection. By converting a 0.1% mutation to a 10–40% mutation (Figure 6) the number of reads required drops sharply to less than 100. Accordingly, the reduction in coverage required to call low level mutations is anticipated to translate to an equivalent reduction in throughput and cost associated with next generation sequencing. Fast-TT-COLD-PCR can be used either alone or combined with other noise-reduction technologies like TAm-Seq (12), error-suppressed sequencing (11), or single molecule-barcoding sequencing, SafeSeq (13), to enable even higher throughput and accuracy.

There was also a need to adapt fast-COLD-PCR to the enrichment also of DNA variations that increase the melting temperature of the amplicon (A:T>G:C or A:T>C:G), which represent ~26% of all reported missense mutations (35). Such mutations could in principle be enriched by using full-COLD-PCR (31) or ice-COLD-PCR (29, 38). However the DNA intermediate hybridization required by these approaches limits the levels of multiplexity that can be achieved in a single-tube reaction as with increasing the number of amplicons the heteroduplex formation decreases. Fast-COLD-PCR does not have the requirement for intermediate hybridization (21, 24, 39) thus very high levels of multiplexity can be achieved (Figures 5 and Supplemental Figure 5). The application of fast-COLD-PCR in the presence of modified bases enables preferential amplification of mutations that increase the melting temperature of the amplicon (A:T>G:C and A:T>C:G).

The strategy followed in this proof-of-principle study, employed co-amplification of groups of amplicons with diverse melting temperatures using a step-up COLD-PCR approach (Figure 1). Within each group, only those amplicons having a Tc contained within a chosen temperature window of ~2°C are enriched for mutations; amplicons with lower melting temperature would be expected to amplify without discriminating for mutant sequences, thus potentially saturating the PCR reaction, limiting the mutation enrichment achieved for the remaining sequences, and generating sequence representation differences for downstream NGS. While the mutation enrichment was successful, an even more efficient approach might have been to group the panels of amplicons within each fast TT-COLD-PCR reaction such that they all have similar melting temperatures and Tc,, thereby enabling all amplicons to enrich for mutations simultaneously. Work in this direction is currently underway.

In summary, we provided proof of principle for a novel approach utilizing fast-COLD-PCR to improving the limit of detection of cancer gene panels. Incorporating fast-TT-COLD-PCR in the sample preparation workflow enables enrichment of mutations prior to performing the traditional Sanger sequencing or prior to initiating NGS-based profiling. This modification provides the ability to treat subclonal mutations as if they were clonal, decreasing the required sequencing depth for reliable detection.

Supplementary Material

supplem_Tables
supplem_fig_1-7

Acknowledgments

This work was supported by National Cancer Institute grants 1 R41 CA180389-01 and 1 R21 CA175542-01A1. The contents of this manuscript do not necessarily represent the official views of the National Cancer Institute.

Footnotes

CONFLICT OF INTEREST STATEMENT: COLD-PCR is a technology owned by the Dana-Farber Cancer Institute and licensed, in part, to Transgenomic Inc. The manuscript has not been published previously and is not being considered concurrently by another journal. All authors and acknowledged contributors have read and approved the manuscript.

REFERENCES

  • 1.Diaz LA, Jr, Williams RT, Wu J, Kinde I, Hecht JR, Berlin J, et al. The molecular evolution of acquired resistance to targeted egfr blockade in colorectal cancers. Nature. 2012;486:537–540. doi: 10.1038/nature11219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Misale S, Yaeger R, Hobor S, Scala E, Janakiraman M, Liska D, et al. Emergence of kras mutations and acquired resistance to anti-egfr therapy in colorectal cancer. Nature. 2012;486:532–536. doi: 10.1038/nature11156. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Dawson SJ, Tsui DW, Murtaza M, Biggs H, Rueda OM, Chin SF, et al. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med. 2013;368:1199–1209. doi: 10.1056/NEJMoa1213261. [DOI] [PubMed] [Google Scholar]
  • 4.Geurts-Giele WR, Dirkx-van der Velden AW, Bartalits NM, Verhoog LC, Hanselaar WE, Dinjens WN. Molecular diagnostics of a single multifocal non-small cell lung cancer case using targeted next generation sequencing. Virchows Arch. 2013;462:249–254. doi: 10.1007/s00428-012-1346-4. [DOI] [PubMed] [Google Scholar]
  • 5.Beadling C, Neff TL, Heinrich MC, Rhodes K, Thornton M, Leamon J, et al. Combining highly multiplexed pcr with semiconductor-based sequencing for rapid cancer genotyping. J Mol Diagn. 2013;15:171–176. doi: 10.1016/j.jmoldx.2012.09.003. [DOI] [PubMed] [Google Scholar]
  • 6.Yousem SA, Dacic S, Nikiforov YE, Nikiforova M. Pulmonary langerhans cell histiocytosis: Profiling of multifocal tumors using next-generation sequencing identifies concordant occurrence of braf v600e mutations. Chest. 2013;143:1679–1684. doi: 10.1378/chest.12-1917. [DOI] [PubMed] [Google Scholar]
  • 7.Tsongalis GJ, Peterson JD, de Abreu FB, Tunkey CD, Gallagher TL, Strausbaugh LD, et al. Routine use of the ion torrent ampliseq cancer hotspot panel for identification of clinically actionable somatic mutations. Clin Chem Lab Med. 2014;52:707–714. doi: 10.1515/cclm-2013-0883. [DOI] [PubMed] [Google Scholar]
  • 8.Johansson H, Isaksson M, Sorqvist EF, Roos F, Stenberg J, Sjoblom T, et al. Targeted resequencing of candidate genes using selector probes. Nucleic Acids Res. 2011;39:e8. doi: 10.1093/nar/gkq1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Milbury CA, Li J, Makrigiorgos GM. Pcr-based methods for the enrichment of minority alleles and mutations. Clin Chem. 2009 doi: 10.1373/clinchem.2008.113035. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Milbury CA, Correll M, Quackenbush J, Rubio R, Makrigiorgos GM. Cold-pcr enrichment of rare cancer mutations prior to targeted amplicon resequencing. Clin Chem. 2012;58:580–589. doi: 10.1373/clinchem.2011.176198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Narayan A, Carriero NJ, Gettinger SN, Kluytenaar J, Kozak KR, Yock TI, et al. Ultrasensitive measurement of hotspot mutations in tumor DNA in blood using error-suppressed multiplexed deep sequencing. Cancer Res. 2012;72:3492–3498. doi: 10.1158/0008-5472.CAN-11-4037. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Forshew T, Murtaza M, Parkinson C, Gale D, Tsui DW, Kaper F, et al. Noninvasive identification and monitoring of cancer mutations by targeted deep sequencing of plasma DNA. Sci Transl Med. 2012;4:136ra68. doi: 10.1126/scitranslmed.3003726. [DOI] [PubMed] [Google Scholar]
  • 13.Kinde I, Wu J, Papadopoulos N, Kinzler KW, Vogelstein B. Detection and quantification of rare mutations with massively parallel sequencing. Proc Natl Acad Sci U S A. 2011;108:9530–9535. doi: 10.1073/pnas.1105422108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hiatt JB, Pritchard CC, Salipante SJ, O'Roak BJ, Shendure J. Single molecule molecular inversion probes for targeted, high-accuracy detection of low-frequency variation. Genome Res. 2013;23:843–854. doi: 10.1101/gr.147686.112. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Schmitt MW, Kennedy SR, Salk JJ, Fox EJ, Hiatt JB, Loeb LA. Detection of ultra-rare mutations by next-generation sequencing. Proc Natl Acad Sci U S A. 2012;109:14508–14513. doi: 10.1073/pnas.1208715109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Newman AM, Bratman SV, To J, Wynne JF, Eclov NC, Modlin LA, et al. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat Med. 2014;20:548–554. doi: 10.1038/nm.3519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Flaherty P, Natsoulis G, Muralidharan O, Winters M, Buenrostro J, Bell J, et al. Ultrasensitive detection of rare mutations using next-generation targeted resequencing. Nucleic Acids Res. 2012;40:e2. doi: 10.1093/nar/gkr861. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Koboldt DC, Chen K, Wylie T, Larson DE, McLellan MD, Mardis ER, et al. Varscan: Variant detection in massively parallel sequencing of individual and pooled samples. Bioinformatics. 2009;25:2283–2285. doi: 10.1093/bioinformatics/btp373. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Bansal V, Libiger O, Torkamani A, Schork NJ. Statistical analysis strategies for association studies involving rare variants. Nat Rev Genet. 2010;11:773–785. doi: 10.1038/nrg2867. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Bidard FC, Weigelt B, Reis-Filho JS. Going with the flow: From circulating tumor cells to DNA. Sci Transl Med. 2013;5:207ps14. doi: 10.1126/scitranslmed.3006305. [DOI] [PubMed] [Google Scholar]
  • 21.Li J, Wang L, Mamon H, Kulke MH, Berbeco R, Makrigiorgos GM. Replacing pcr with cold-pcr enriches variant DNA sequences and redefines the sensitivity of genetic testing. Nat Med. 2008;14:579–584. doi: 10.1038/nm1708. [DOI] [PubMed] [Google Scholar]
  • 22.Castellanos-Rizaldos E, Liu P, Milbury CA, Guha M, Brisci A, Cremonesi L, et al. Temperature-tolerant cold-pcr reduces temperature stringency and enables robust mutation enrichment. Clinical Chemistry. 2012;58:1130–1138. doi: 10.1373/clinchem.2012.183095. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Carotenuto P, Roma C, Cozzolino S, Fenizia F, Rachiglio AM, Tatangelo F, et al. Detection of kras mutations in colorectal cancer with fast cold-pcr. Int J Oncol. 2012;40:378–384. doi: 10.3892/ijo.2011.1221. [DOI] [PubMed] [Google Scholar]
  • 24.Kristensen LS, Daugaard IL, Christensen M, Hamilton-Dutoit S, Hager H, Hansen LL. Increased sensitivity of kras mutation detection by high-resolution melting analysis of cold-pcr products. Hum Mutat. 2010;31:1366–1373. doi: 10.1002/humu.21358. [DOI] [PubMed] [Google Scholar]
  • 25.Pritchard CC, Akagi L, Reddy PL, Joseph L, Tait JF. Cold-pcr enhanced melting curve analysis improves diagnostic accuracy for kras mutations in colorectal carcinoma. BMC Clin Pathol. 2010;10:6. doi: 10.1186/1472-6890-10-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Milbury CA, Li J, Makrigiorgos GM. Cold-pcr-enhanced high-resolution melting enables rapid and selective identification of low-level unknown mutations. Clin Chem. 2009;55:2130–2143. doi: 10.1373/clinchem.2009.131029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Dwight Z, Palais R, Wittwer CT. Umelt: Prediction of high-resolution melting curves and dynamic melting profiles of pcr products in a rich web application. Bioinformatics. 2011;27:1019–1020. doi: 10.1093/bioinformatics/btr065. [DOI] [PubMed] [Google Scholar]
  • 28.Guha M, Castellanos-Rizaldos E, Liu P, Mamon H, Makrigiorgos GM. Differential strand separation at critical temperature: A minimally disruptive enrichment method for low-abundance unknown DNA mutations. Nucleic Acids Res. 2013;41:e50. doi: 10.1093/nar/gks1250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Milbury CA, Li J, Makrigiorgos GM. Ice-cold-pcr enables rapid amplification and robust enrichment for low-abundance unknown DNA mutations. Nucleic Acids Research. 2011;39:e2. doi: 10.1093/nar/gkq899. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Mamon H, Hader C, Li J, Wang L, Kulke M, Amicarelli G, et al. Preferential amplification of apoptotic DNA from plasma: Potential for enhancing detection of minor DNA alterations in circulating DNA. Clinical Chemistry. 2008;54:1582–1584. doi: 10.1373/clinchem.2008.104612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Li J, Makrigiorgos GM. Cold-pcr: A new platform for highly improved mutation detection in cancer and genetic testing. Biochemical Society Transactions. 2009;37:427–432. doi: 10.1042/BST0370427. [DOI] [PubMed] [Google Scholar]
  • 32.Li J, Wang L, Janne PA, Makrigiorgos GM. Coamplification at lower denaturation temperature-pcr increases mutation-detection selectivity of taqman-based real-time pcr. Clin Chem. 2009;55:748–756. doi: 10.1373/clinchem.2008.113381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. Integrative genomics viewer. Nat Biotechnol. 2011;29:24–26. doi: 10.1038/nbt.1754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Suspene R, Renard M, Henry M, Guetard D, Puyraimond-Zemmour D, Billecocq A, et al. Inversing the natural hydrogen bonding rule to selectively amplify gc-rich adar-edited rnas. Nucleic Acids Res. 2008;36:e72. doi: 10.1093/nar/gkn295. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Greenman C, Stephens P, Smith R, Dalgliesh GL, Hunter C, Bignell G, et al. Patterns of somatic mutation in human cancer genomes. Nature. 2007;446:153–158. doi: 10.1038/nature05610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Milbury CA, Li J, Liu P, Makrigiorgos GM. Cold-pcr: Improving the sensitivity of molecular diagnostics assays. Expert Rev Mol Diagn. 2011;11:159–169. doi: 10.1586/erm.10.115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Murphy DM, Bejar R, Stevenson K, Neuberg D, Shi Y, Cubrich C, et al. Nras mutations with low allele burden have independent prognostic significance for patients with lower risk myelodysplastic syndromes. Leukemia. 2013;27:2077–2081. doi: 10.1038/leu.2013.160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.How Kit A, Mazaleyrat N, Daunay A, Nielsen HM, Terris B, Tost J. Sensitive detection of kras mutations using enhanced-ice-cold-pcr mutation enrichment and direct sequence identification. Hum Mutat. 2013;34:1568–1580. doi: 10.1002/humu.22427. [DOI] [PubMed] [Google Scholar]
  • 39.Luthra R, Zuo Z. Cold-pcr finds hot application in mutation analysis. Clinical chemistry. 2009;55:2077–2078. doi: 10.1373/clinchem.2009.136143. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplem_Tables
NIHMS640207-supplement-supplem_Tables.pdf (119.8KB, pdf)
supplem_fig_1-7
NIHMS640207-supplement-supplem_fig_1-7.pdf (513.8KB, pdf)

RESOURCES